Keywords

1 Introduction and Context

Floodplains are important and unique ecosystems along lowland rivers and process large fluxes of energy and materials transferred from upstream sources (Hughes 1997). The dynamic nature of alluvial rivers is a function of flow and sediment regime s interacting within geologic, physiographic, and land use and land-cover controls (Ward and Stanford 1995). Economic development and expansion has strongly affected river and floodplain hydrology and associated ecosystems in many temperate and tropical areas to a larger degree than most other natural systems (Bayley 1995). Embankments and dams have led to rapid and drastic changes to floodplain and riverine ecosystems, impairing biota and abiotic processes (Hintz 2011; Ward and Stanford 1995; Clawson et al. 2001). From the standpoint of this substantially altered condition, the adaptation of floodplain ecosystems to the new context of climate change and greater human pressure represents a substantial challenge for floodplain managers and society.

Long-term adaptation consists of a series of complex processes involving climate change , river regulation and community responses (Martens and Chang 2010; Jenkins et al. 2012). Some impending threats acting upon these fragile ecosystems include potential increases in drought frequency and duration, temperature increases, sea-level rise and changes in rainfall regimes. From a historical perspective over the past century, anthropogenic intervention in the Danube basin has been the most important influence on floodplain dynamics. The potential to modify the floodplain to mitigate against further climate change and human pressure is constrained by the narrower embanked floodplain and modified hydrologic regime . The presence of dams has been associated with a reduction in sediment loads, as well as increased channel velocity . Combined, the effect has been to increase river bank erosion . The EU Water Framework Directive represents a fundamental guide for conducting floodplain restoration projects. According to Moss and Monstadt (2008), important restoration schemas were already undertaken on the Rhine, Elbe, Garonne and Long Eau Rivers.

The purpose of this study is to document the history of human impacts to the lower Danube floodplain in the context of water resource control and floodplain management . By emphasizing the interaction of fluvial processes with human development, the study represents an important baseline to management and restoration efforts—along the Danube as well as other large European rivers.

The Danube is among Earth’s most international rivers and is the largest fluvial system within the European Union in terms of length, drainage area, discharge , and sediment load (6470 m3/s and 1555 kg/s, respectively; McCarney-Castle 2012). As it flows over 2870 km from its headwaters in Germany and Switzerland to its delta at the western Black Sea coast, the Danube River flows within 10 nations and its basin drains 19 nations within an area of 817,000 km2. The variation in river stage is large (10.5 m), among the highest stage variations of temperate zone rivers (Vidraşcu 1921), and historically was associated with a distinctive floodplain pulse and lateral hydrologic connectivity . Downstream of Iron Gates Gorge (Fig. 1), an extensive natural floodplain bordered the Danube along its lower course, which included numerous lakes rich in fish and other forms of aquatic wildlife. Historically, for example, the Balta Brăilei floodplain region upstream of the delta was associated with one of the most productive freshwater fisheries in Europe (Antipa 1910). After the 1960s, however, an extensive reclamation program converted the floodplain wetlands and lakes into predominantly an agricultural region, and is considered the most devastating and abrupt anthropogenic transformation of a fluvial wetland in post-war Europe (Botnariuc and Vădineanu 1982).

Like many large temperate zone rivers, the sediment load of the Danube River has been substantially reduced because of dam and reservoir construction. The Danube basin has many hundreds of dams and reservoirs, which were constructed primarily after the World War II . Over 150 dams were constructed within the Romanian portions of the basin, which have reservoirs that can store up to 22 billion m3 of water. Several dams were built on local rivers in Romania (Fig. 3) and considerably reduced downstream sediment loads by trapping sediments within the reservoirs. The Jiu, Olt, Argeş, Ialomiţa and Siret rivers previously provided about a third of the Danube’s annual sediment load. Because of the erodible loess deposits in upstream basins, the sediment discharge was naturally high within this section of the basin, and increased by ~ 37 % between Orşova and the Danube delta . This increase in sediment load was associated with only a 16 % in stream flow (Mihăilescu 1969). Along these tributaries, many small dams were built after 1950 and as a result, available measurements indicate considerable sediment load reductions along the Danube, including ~ 69 % reduction from the Jiu River, ~ 67 % reduction from the Argeş River, and a ~ 48 % reduction from the Siret and Prut Rivers (Rădoane 2008).

In addition to larger dams within Romania , during about a 30 year period following World War II over 600 small dams were also built on Danube tributaries in Bulgaria . These dams considerably reduced sediment discharge to the lower Danube, from 4.4 million t/year to only 0.4 million t/year (Levashova et al. 2004). Overall, Danube tributaries are currently contributing ~ 60 % less suspended sediment than under pre-dam conditions (McCarney-Castle 2012).

The Danube River course comprises 56 permanent main-stem reservoirs, amounting to 4.8 billion m3 of water storage (Stanciu et al. 2008). The construction of the Iron Gates I and II Dams (Fig. 1) was a major contributor to the sediment reduction in the lowermost reaches of the Danube. Combined, the sediment load reduced by 53 % at the entry to the delta (1846 kg/s between 1840 and 1970 and 962 kg/s between 1971 and 2000) (Bondar 2008).

Fig. 1
figure 1

The lower course of the Danube River with its main tributaries together with the floodplain segmented into morphological sectors. The map reflects the situation prior to 1960. P1 to P6 represents the location of profiles from Fig. 5

Full size image

2 Motivation

The first debates regarding the agricultural development of the Danube floodplain occurred early in the twentieth century with proposals from the Romanian polymath Grigore Antipa (1867–1944) and his collaborator, engineer G. Vidrascu. Antipa’s ideas were well advanced for their time: ecologically-oriented and grounded in the flood pulse concept with minor negative effects to be expected on the fluvial ecosystem . In contrast, the reclamation of wetlands for agriculture after 1950 during the Communist period completely transformed the floodplain. Consequently, its current heavily degraded state requires massive restoration efforts (Vădineanu 2001).

Prior to 1950’s, the river dynamics greatly influenced the livelihoods of floodplain inhabitants along the lower Danube. Rare, exceptional hydrologic events, whether floods, droughts, or river course changes left deep imprints in the collective memory of residents of wetland communities, leading to a rich collective knowledge of the fluvial environment. The subsequent transformation of the region, however, and the new sense of security brought in by reclamation of wetlands caused the riparian communities to turn their thoughts from nature. Engineering the river has fulfilled an insidious double role of flood protection , but also to irreversibly isolate floodplain residents from their natural environment. Protection structures have thus become a symbol not only for the artificial state of the environment, but also of the state of the human spirit. Future efforts of floodplain restoration and development should include not only a return to the idea of a minimal developmental footprint as originally proposed by Antipa and Vidrascu, but also to a reconstruction of the cultural dimension of the way of life for floodplain inhabitants. Our contribution to these recovery efforts is the present paper that surveys and discusses the morphology of the floodplain under natural and anthropogenic conditions.

3 Mapping Methods

Following the Treaty of Berlin in 1878 that recognized its de facto independence, Romania began to systematically collect spatial data pertaining to topography and hydrolologic features as it strove to map its entire territory. These efforts between 1880 and 1899 enabled cartographers from Romanian Army’s Institute for Cartography to produce maps of the entire lower Danube valley. The spatial data were drafted to maps using a Lambert-Cholesky projection, and is further described in Bartos et al. (2007). A transition to a Gauss-Kruger system in 1951 required considerable effort to update the cartography for the entire nation at a 1:25,000 scale . The latest updates to these maps occurred in the 1980s and utilized new field measurements and aerial images. Since the 1980s satellite and aerial images have become available, and in this study are utilized for the period of analysis spanning from 2005 to 2008. All spatial data were processed and transformed into a unique Stereo-70 projection, the official projection used in Romania.

4 Background and Terminology

The all-encompassing term “Danube floodplain” includes the entire surface subject to flooding from the modern hydrologic regime , as well as higher alluvial terraces (Posea 2005). The common Romanian term for this region is baltă (loosely translated as “the water realm” but meaning interchangeably “pond”, “wetland”, “lake”, ”river”; (Conea and Badea 2006)), which was previously used in the local scientific jargon (e.g., de Martonne 1902; Munteanu-Murgoci 1907). The term lac (English: lake) is not used in the local informal language. To avoid confusion between its multiple meanings, Antipa introduced the notion of flooding zone, a phrase that was later adopted by Vidraşcu. The purpose was to distinguish between the larger geomorphic units composed of low terrains near the river and its components such as the riverbed or the floodplain proper. After the communist reclamation program, the term baltă faded away from the local collective memory being replaced by luncă (English: floodplain), which was preferentially employed in official documents. The new term was more convenient for communist authorities as it signified a new reality, a transformed space, whereas the popular name referred to a past quickly forgotten.

The Danube floodplain varies in extent along it lower course. Downstream of the Iron Gates I Dam, most islets and the surrounding small floodplain areas located within the upstream gorge were completely inundated (Fig. 2). Immediately downstream of Iron Gates Gorge , which is considered the formal upstream boundary to the lower Danube River , the floodplain is narrow and sporadic. At Ostrovul Mare the floodplain begins to broaden preferentially on the left bank (Romania ), and here the floodplain width varies between ~ 200 m near Calafat and ~ 30 km in Balta Brăilei. On the right bank, in Bulgaria , the floodplain is a narrow fragmented strip that was largely embanked before World War II (55,000 ha reported by Hâncu and Dan 2008; 88,000 ha reported by Ioaniţoaia 2007). In Romania, along the 993.5 km long course of the Danube from the Nera River tributary down to Ceatal Izmail at the entrance to the Danube delta , the Danube floodplain extends over 573,000 ha. About 75 % of the floodplain is currently embanked. Within this embanked region, the surface covered by water amounts to 11,143 ha and is dedicated to fisheries, while the remaining area is used for agriculture (Maria 2008). The total length of the embankments along the main course of the Danube is 3520 km, of which 1158 km are located in Romania.

Fig. 2
figure 2

The lower course of the Danube floodplain with its main floodplain enclosures

Full size image
Fig. 3
figure 3

Water (Ql) and sediment (Qs) discharges at the Drobeta Turnu Severin (DTS) and Brăila gauges between 1960 and 2007 (in solid and dashed lines respectively). Note changes in discharge after the Iron Gates dams construction

Full size image

Flood basins including lakes and marshes are the most important parts of the floodplain in terms of habitat and fauna diversity. They previously served as storage basins and represented safety valves during floods, a feature noted in studies at the beginning of the last century (Vidrașcu 1921; Antipa 1921). Along the Romanian side of the floodplain, before reclamation, there were ca. 2050 floodplain basins. Following drainage and engineering works they were reduced in surface and many disappeared altogether. In natural conditions, three types of lake environments occurred (Fig. 4): (1) between Severin and Calafat only partly infilled single cut-off oxbow lake s were present, (2) between Calafat and Giurgiu large single lake basins were more common, but were rarely associated in complexes, and (3) downstream of Giurgiu, lake complexes become instead common (Posea 2005).

Fig. 4
figure 4

Morphology of Danube floodplain lakes. a oxbow lake s, b single lakes infrequently associated in complexes, c lake complexes

Full size image
Fig. 5
figure 5

Cross-sectional profiles of the Danube floodplain (see Fig. 1 for locations)

Full size image

5 Danube Floodplain Reaches

Based on geomorphological features the lower Danube floodplain was classified into several reaches (Vidraşcu 1921; Mihăilescu 1969; Posea et al. 1974, 2005). The absolute elevation of the floodplain is of 65–70 m before the Iron Gate dam, 35 m at Drobeta Turnu Severin, 20–22 m at the Olt confluence, 15 m at the Argeş confluence, and 10 m at Călăraşi and 5 m at Brăila. The elevation gradually decreases and reaches 2 m at Ceatal Izmail. The 134 km long gorge, or the upper sector, extends from Baziaş to Gura Văii and is characterized by small floodplain patches (Fig. 5 P1). Following the construction of the Iron Gates I Dam, most floodplain patches were inundated. The Corbul Islet, a 4 km wide sector from an artificially straightened meander , is the largest floodplain patch downstream of Iron Gates II and comprises a 1700 ha enclosure separated by a 5.7 km long embankment. Another enclosure, of only 264 ha, was built on Ostrovul Mare.

Downstream of Ostrovul Mare, the floodplain widens and reaches a maximum span of approximately 30 km, primarily on the left bank down to Brăila (Fig. 1). The Oltenia terraced plain sector (256 km) and the Burnaz Danube Corridor (227 km) are separated by the Olt valley. The first is characterized by a sequence of terraced plains with sand dunes between Batoţi and Bechet, followed by a 50–150 m escarpment of the Pre-Balcanic Plateau (Fig. 5 P2, P3). The Burnaz Corridor, east of the Olt River appears as a long valley with particularly active geomorphic processes and floodplain bottlenecks controlled by alluvial fans at tributary junctions (Fig. 5 P4). In these two central sections (Oltenia and Burnaz), the average floodplain width ranges from 5 to 6 km.

Farther downstream, the Danube branches repeatedly between Călărași and Brăila, and consists of two distinct sections: Balta Ialomiței and Brăilei (Fig. 6). The valley in these sections is asymmetric with the left bank developed along the low altitude Romanian Plain and the right bank along the Dobrogea Plateau with an elevation of more than 100 m. Balta Ialomiței extends over 130 km along the course between Călărași and Giurgeni-Vadu Oii, with a maximum width of 18 km. The crescent-like sector, concave toward the Romanian Plain, changes in elevation from 10 to 11 to 8 m from the south to its northern extremity (Fig. 5 P5). The presence of lakes or fluvial limans at tributary junctions on the Dobrogea Plateau side is a unique characteristic of this section. Downstream, Balta Brăilei is only 70 km long but widens to extend over ~ 30 km (Fig. 5 P6). The sequential branching of the river, specific to the Călăraşi-Brăila sector, is indicative of a decrease in slope and sediment transport capacity (Fig. 6). The Small Islet of Brăila, with a surface of 17,529 ha located in the western part of Balta Brăilei, is the only floodplain area along the lower Danube valley subject to the natural flood regime (Toader 2005). In 2001 this section was declared a RAMSAR Wetland Site, and designated as a nature preserve in 2003.

Downstream of Brăila, the Danube valley becomes more symmetrical with its banks at low elevations. This 90 km long reach along the lower sector is also known as Balta Isaccei. Here, the river flows through a single channel until the first bifurcation at Ceatal Izmail, the apex of the Danube delta .

Fig. 6
figure 6

Danube floodplain between Călăraşi and Brăila

Full size image

6 From Natural to an Engineered Floodplain

At the beginning of the twentieth century, two different conceptual models emerged for the use of the Danube floodplain. The first model championed by engineer Anghel Saligny and agronomist Gheorghe Ionescu-Sisești (1933), proposed a plan to drain the floodplain for agricultural development. Alternatively, a second model proposed by the polymath biologist Grigore Antipa (1910, 1913, 1928) and engineer Ion Vidrașcu (1915, 1921) was in favour of preserving the natural flood regime and floodplain features to develop fish farms and animal husbandry. After 1960, the agricultural model was implemented through extensive embankments and drainage works along lower Danube (Bondar 2008). As a result, over the next two decades the natural hydrologic and geomorphic regime was largely eliminated to make way for intensive agriculture . Despite their economic benefits, embankments along the river resulted in a narrowing of the channel bed with a direct impact on the hydrologic regime . After 1990, most of the floodplain returned to private ownership , renewing the debate of how to develop the floodplain. The conservative model promoted the continuation of agricultural activities, while an ecological model advocated a return to the natural hydrogeomorphic regime through the removal of engineering structures.

7 Natural Evolution: Flooding and Geomorphic Processes Before 1950

Under natural conditions when the Danube overtopped its banks, the width of the river increased ~ 4 to 12 times (Antipa 1910). The surface covered by lakes was approximately 14 % of the floodplain (Antipa 1910), and the floodplain basins accommodated the surplus of fluvial waters during floods. In contrast to the modern engineering position which negatively views floods from the perspective of risk, floods were considered differently in the past. Until the early twentieth century floods were appreciated as natural gifts, which today would be analogous to viewing floods as ecosystem services . Years of major floods, when the Danube inundated the floodplain, were acknowledged as the most productive for fisheries. Before the construction of the embankments seasonal high discharge events mainly did not overtop the natural levees , with the discharge wave requiring 5–6 days to travel between Turnu Severin and Galaţi as in e.g., 1895, 1890 (Vidraşcu 1921). During high discharge years the flood wave was delayed up to 30 days because of flood water storage across the floodplain (e.g., 1889, 1897). Between 1921 and 1960, the maximum discharge generally occurred in April and May, in the Baziaş-Giurgiu sector. Downstream of Giurgiu the maximum discharge occurred in June and July, because of the slow propagation and strong attenuation of the flood wave by the floodplain geomorphology (Mihăilescu 1969). The important point of these examples is that the floodplain played an important role in attenuating the effects of floods. In 1897, for example, water storage in Balta Brăilei alone reached up to 5.5 billion m3 of the total of 24 billion m3 of water stored along the entire floodplain (Antipa 1921).

According to Vidrascu (1915), rather than infrequent extreme floods, a moderate hydrologic regime over long periods of time is the dominant force in the evolution of large river valley geomorphology . Floodplain characterization, therefore, should consider the geomorphic evolution of a river system over long time scales to understand the behaviour of fluvial system. To that effect, the Romanian engineer Ion Vidrașcu used hypsometry and hydrological characteristics to introduce a novel unit, the hydrodegree, which was locally defined by the maximum flood height. The hydrodegrees enable the following floodplain elements to be differentiated:

  1. 1.

    Natural levees with the highest elevations, overtopped only during exceptional floods. The elevation of the levee corresponds to an average value of discharge in the spring months of 7 hydrodegrees. For instance, at Dunăreni (Fig. 6) the 7 hydrodegrees corresponds to 5.10 m local stage and the elevation of the levee is between 4.6 and 5.2 m.

  2. 2.

    Lakes are the geomorphic units with the lowest elevations thus 0 hydro degrees. Our GIS reconstruction shows that in the natural regime, lakes represented 19.2 % of the entire surface between Călăraşi and Brăila compared to 2.9 % in 2005.

  3. 3.

    Floodplain marsh with Typha and reeds is the third geounit, having intermediate elevations between 0 and 7 hydrodegrees.

In natural conditions floodplain and lakes exchanged water with the river via a secondary natural stream network. While during the early twentieth century the density of this network between Călăraşi and Brăila amounted to 0.92 km/km2, the current channel density has increased to 1.66 km/km2 because of artificial channelling for drainage and irrigation . Many of these artificial channels are currently in an advanced state of infilling and have lost their efficiency. Natural stream channels reactivate only during large floods, such as those in 2005 and 2006, in areas where the protective embankments failed (Fig. 2). During normal spring floods, the river used to overflow into lakes, often covering the entire floodplain with water. When the Danube level decreased, the system reversed with water movement from lakes to the river. Discharge into the marshes reached 5–6 m3/s. The discharge rate for the Filipoiu marsh, located in Balta Brăilei, was reported to be 450 m3/s towards the Măcin branch in 1906 (Vidrascu 1915). The return flow was active by definition until 3 hydrodegrees, which corresponds to between + 2 and + 2.5 m in the Brăila-Hârșova sector. In 1897, a year of major floods, the floodplain between Giurgiu and Brăila had a storage capacity of 5–6 % of the entire Danube discharge. The sector between Brăila and Hârșova, although covering a region three times smaller than Danube Delta, had a storage capacity of 80 % of the entire delta (Vidrașcu 1915). The importance of the Balta Ialomiţei area in flood mitigation is clearly shown in the analyses of the discharge rate at Vadu Oii, where the average discharge was 231 m3/s lower than upstream at Silistra (Tufescu 1974).

These data emphasize the importance of the floodplain in relation to the hydrologic regime , and further reveals the existence of feedback processes between lakes, marshes and river subsystems. The rotating polder system of Antipa, as opposed to the agricultural model, proposed a modern—integrated—development vision, with minor changes to fluvial ecosystems. During dry years, which were more damaging than floods to agriculture , the floodplain could be utilized as pasture for a source of revenue. During wet years fisheries revenue increased, compensating for the lost revenue from pasture. The relationship between the extent of the flood prone areas and fish production was well-known at the beginning of the twentieth century. The Communist regime brought new technology and a vision which strongly strayed from the natural rhythm, imposing human control to “conquer nature”.

8 Anthropogenic Evolution: Flooding and Geomorphic Processes After 1950

The first engineering works were initiated at the end of the nineteenth Century, but by far the most extensive changes occurred after 1960. Between 1904 and 1906, floodplain was reclaimed at Chirnogi (1058 ha), Simoiu-Mânăstirea (334 ha) and Luciu Giurgeni (3150 ha). The floodplain was embanked during successive droughts occurring between 1904 and 1916. Large sections of embankments were built at Spanţov (1780 ha) between 1906 and 1908. This is also the location for the first agricultural research station on embanked floodplains (Ioaniţoaia 2007). The total surface of these embankments amounted to 23,370 ha 1928, but the agricultural production was below expectations because of the lack of irrigation and flood protection . The approach, based on the natural flood pulse and oriented toward developing fisheries, was the preferred alternative to be implemented by the Administration of Danube Fisheries and Floodable Land Improvement (Administraţia Pescăriilor şi Ameliorării regiunii Inundabile a Dunării—PARID). Subsequently, various land use authorities took charge of the works in the Danube floodplain (Ioaniţoaia 2007).

By 1962 101,000 ha of floodplain embankments were constructed, including 18 complete enclosures. The most intense period of floodplain development occurred between 1963 and 1971 (Fig. 2) when 289,000 ha of embankments were constructed with 24 new enclosures. Between 1971 and 1990, 14 new enclosures were constructed with a total surface of 41,800 ha. At present, 56 embanked enclosures subdivide the lower Danube floodplain in Romania . These enclosures cover 431,763 ha, with 55 % located on the left bank, 12 % on the right bank and 33 % on islets. Protective embankments with a total length of 1158 km are located on the floodplain with 619 km on the left bank, 175 km on the right bank and 31 km on islets (Ioaniţoaia 2007). As of 2007 the floodplain had the following land use distribution: arable lands 70.8 %, forests 10.3 %, fishery 3.5 %, reed processing 0.32 %, residential 1.49 %, transportation 6.7 % and unused 5.37 %.

The embankment works are described rather ambiguously in the literature, frequently using redundant information. The eulogistic tone, specific to the Communist era, is used when presenting the engineering works as a conquest of nature, vital for the economic development of Romania . This rhetoric served as a justification for actions over 50 years, each time invoking the need for increased agricultural production. The scientific literature, however, acknowledges that embankments caused a significant increase in the river stage during extensive discharge events in 1965, 1970, 1985, 2005 and 2006 when the floodplain was preferentially flooded upstream of Bechet (Vişinescu and Bularda 2008). Seven enclosures, for example, were flooded in 2006 covering 72,700 ha. In addition, two enclosures covering 15,165 ha were deliberately flooded to lower the stage of the floodwave (i.e., Borcea-Răul and Făcăeni-Vlădeni) (Fig. 2 in bold).

Embankment and drainage activities have completely altered floodplain geomorphic processes. The narrowing of the river bed and artificial levee (dike ) construction resulted in an increase in streamflow, causing increased lateral erosion and river bed incision . As a result, the discharge increased by up to ~ 2500 m3/s and the high water stage along the Danube was raised by 0.5 to 1.20 m (Mihăilescu 1969). These changes led to an acceleration of geomorphic processes that is best illustrated by the dynamics of fluvial islets (see section below). The embankments were built to withstand a 100-year flood upstream of Călăraşi, but with 5–10 % protection downstream (Ioaniţoaia 2007). The water volume stored behind these embankments amounts to ~ 7 billion m3 along the Iron Gates II-Călăraşi sector and approximately 11 m3 between Călăraşi and Isaccea. The embankment height varies between 3.5 and 4 m. The distance to the river valley varies from 150 to 200 m to 300–400 m.

In addition to floodplain development, agricultural lands in southern Romania rely upon Danube River water for irrigation . Prior to 1989 irrigated lands amounted to 2.3 Mha, requiring not only impressive amounts of water but also high energy consumption to drive irrigation pumps. Sixty-nine percent of the total lands were on terraces , at 60–70 m elevation. If a minimum of 2500 m3/s of Danube discharge is conserved to maintain navigation , only 45 % of Danube’s water was available for irrigation and other uses (Stanciu et al. 2008). These data indicate the extreme pressure put on Danube River, which resulted not only in major imbalances in the overall fluvial ecosystem , but also proved to be economically unsuccessful.

The 1158 km of embankments , the irrigation system for 418,000 ha, and the irrigation systems serving 224,000 ha of land are an investment of approximately 2,200 million euros, i.e. 5,250 euros per hectare. Adding the agricultural land preparation works, deforestation, reed removal, preliminary dewatering, modeling , movable and immovable assets of the 400 agricultural farms, buildings in private ownership and other infrastructure works and assets, the total lands and works under protection are estimated to value approximately 8.8 billion euros. (Maria 2008)

9 Floodplain Embankment Effects

Because of the scale (size) of the lower Danube such an intensive system of embankments had a significant impact to the fluvial ecosystems. These impacts included a reduction of fish spawning habitat and isolation of fish populations, decreased nutrient retention capacity, floodplain drought, increased soil salinity, reduction in water exchange with the river and within the floodplain, major changes in the structure and composition of vegetation, and the destruction of the last remaining natural floodplain forests in Europe (Antipa 1921; Vădineanu 2001; Iordache 2005).

In addition to the aforementioned internally imposed anthropogenic impacts, indirect external forcing is also important to understanding the modern hydrogeomorphic regime of the lower Danube. Climate change over the last century coupled with engineering works in the upper and central Danube basin resulted in a major change in discharge regime to the lower basin. In the Baziaş sector, for example, the average increase in water discharge increased by ~ 1,200 m3/s, resulting in 40–50 cm increase in river stage along the lower Danube. In addition, the base level of the Black Sea increased at Sulina by approximately 35 cm, resulting in an increase in river stage upstream to Brăila (Mihailovici et al. 2006). These changes likely contributed to the 2006 flood being the largest in over a hundred and 50 years, since the extreme flood of 1840. At Baziaş the highest discharge reached 15,082 m3/s, while at Isaccea the estimated discharge reached 17,700 m3/s (Şerban 2006). These exceptionally high discharges led to stages up to 60 cm higher than previously reported. Uncontrolled embankment breaks occurred at Rast, Bechet, Spanţov, Oltina and Ostrov and forced authorities to undertake controlled breaches at Borcea-Răul and Făcăeni-Vlădeni (Moraru 2007). These breaches resulted in a cumulative decrease in stage by 28 cm, although not enough to prevent downstream flooding of low-lying settlements (Şerban 2006; Ioaniţoaia 2007).

Fig. 7
figure 7

Anthropogenic impact in the Călăraşi-Brăila sector

Full size image

Compared to the natural floodplain, embankments resulted in radical changes in floodplain land use. The arable land surface dramatically increased from 88,000 to 315,713 ha, an increase of 359 %. As a result, the most significant decline is observed in the area covered by forest, which decreased by 93 % from 95,000 to 6269 ha. In addition, other land cover and land use types also underwent considerable reductions, including lake area (− 80.1 %), as well as pastures and grasslands (− 85.5 %). The impact was not homogenous, with higher values in the Călăraşi-Brăila sector, where the lake area decreased by 84.6 % (Fig. 7). Engineering works also led to an increase in other activities which were previously under-represented. The surface area of rice crops, for example, increased to 42,126 ha, while vineyards and orchards covered 2919 ha (Maria 2008). Finally, the general degradation resulting from drainage works and embankments also led to a simplification and a loss of connection to the toponymy of the floodplain (Conea and Badea 2006).

10 Impact of Human Intervention on Fluvial Islets

The occurrence of fluvial islets and their morphometry varies considerably along the 993.5 km length of the Danube, as the river crosses various larger-scale physiographic and geomorphic units. In the upper sector, the river passes through a mountainous region and then a plateau-piedmont region (Mehedinţi Plateau and Getic Piedmont). The asymmetric plain and the Pre-Balcanic Plateau provide different fluvial morphodynamic conditions in the downstream sectors. These conditions are the cause of variations in the riverbed and fluvial islets morphology.

The number of fluvial islets from Baziaş to Isaccea remained largely constant over the last century, varying from 270 in 1920 to 263 in 1980. However, over the last three decades their number dropped by almost 15 % (224 in 2008). The total surface area of the fluvial islets decreased constantly, from 379 km2 (1920), to 341 km2 (1980), to 315 km2 (2008). In order to analyse the spatial distribution of these islets, a density index was calculated in hectares/kilometer (ha/km) of river length. The only sector showing relative stability is the Burnaz Corridor, but in all the other sectors the density decreased over time (Fig. 8). While the formation of the Iron Gates reservoirs explains the reduction in the upper sector, the decrease in sediment quantity may explain the evolution of the other sectors. The highest values of the density index in natural conditions occurred along the Balta Ialomitei and Balta Braila (74.1 ha/km in 1920), while the lowest values were typical for the Lower Sector toward the delta (3.5 ha/km in 1920). River bifurcation along the widest sector of the valley explains these values between Călăraşi and Brăila.

Fig. 8
figure 8

Evolution of the fluvial islet number, density (ha/km) and average surface area (dashed line, in km2)

Full size image

Over the course of about a century (1920–2008) only 6 km2 of fluvial islets disappeared completely. Importantly, with embankment and enclosure the fluvial islets became assimilated into the floodplain, although there was spatial variability along the Danube. The largest average surface area for islets (in km2) occurs in the Bălţi sector (1.7 in 1920; 1.81 in 1980; 1.5 in 2008), followed by the upper sector (1.5; 1.18; 3.08). The change from fluvial to lake regime led to the disappearance of several islets and the coalescence of other islets in the context of a sharp decrease in number, from 38 to 34 to 9. This explains the high average value of surface area: 3.08 (in 2008) in the upper course. An average value for the entire Danube sector reflects a reduction of the average surface area, from 1.27 km2 (1920) to 1.08 (1980), followed by a subsequent increase to 1.46 (2008). This overall decrease, which occurred as a result of intensive improvement works in the floodplain, will likely be reversed as the system attains a new balance.

11 Evolution of River Banks

River bank erosion results in channel bed widening. A negative sedimentary balance occurred over the entire lower Danube. This drove an erosional phase in the fluvial regime, resulting in an average loss of 29.2 ha/km. Only in the Upper Sector the remaining lakes store sufficient water during floods. Downstream of Ostrovul Mare, however, embankments maintain a constant discharge cross-section and increased flood velocities result in channel bank and floodplain erosion (Fig. 9). The only sector where sediment accumulation exceeds erosion is the Balti sector, which explains why navigation on the Old Danube course has become increasingly challenging along this reach. Further, the presence of an underwater rock escarpment in the Izvoarele (Pârjoaia) area led to a discharge deviation from the Old Danube toward Rău Branch, and further to Borcea Branch (Fig. 6). It is estimated that 80 % of this discharge is lost as a result of this obstruction (Ministerul Transporturilor 2005). Given that the Old Danube is the main branch for navigation, several problems occur: the river depth is usually below 2 m and dredging activities are needed every year. Dredging increased from 300,000 to 700,000 m3/year with no corresponding improvement in efficiency. Several attempts to remove the Pârjoaia rock proved ineffective. A new plan to improve navigation conditions is now underway (Ministry of Transportation, Construction and Tourism in Romania 2005). Navigation toward the Danube–Black Sea Canal is thus affected because the alternative route will be 105 km long. To maintain an optimal depth for the Old Danube branch new engineering works have begun at the entrance of the Bala Branch to decrease the discharge and strengthen the upstream banks. Similar works will close the secondary channels between fluvial islets and Dobrogea to redirect the water to the navigable course. Over time, some islets will become part of the river banks, as a result of sedimentation on the closed branches at critical points—Epurașu, Seica Islet (Dunăreni-Mârleanu), Fermecatu Islet, Cochirleni etc.

Fig. 9
figure 9

Intensity of accretion and erosion along the Danube by sector (ha/km) between 1920 and 2008

Full size image

12 Current State of the Floodplain

Political changes after 1989 led to a new social and economic context for the Danube floodplain. A large part of the state-owned enclosures returned to their former owners. Many irrigation canals were not maintained at production standards because high maintenance costs led to a chaotic exploitation in an attempt to obtain easy agricultural profit without investments. The return to an old alternative pits the new owners, who are in favour of maintaining the agricultural terrains for profit, against environmental groups who favour returning the floodplain to natural conditions. Large-scale restoration was rejected as economically inefficient by all Romanian governments after 1990. Restoration can easily be achieved, however, starting with small embanked areas that were abandoned by farmers because of their economic inefficiency caused by salinization and waterlogging.

The European Union Strategy for Danube Region (EUSDR) represents a unitary response to all challenges that affect the entire river basin. This macro-regional strategy, adopted in 2010, was developed by the European Commission in order to coordinate the existing policies and plans across the Danube. At the European level, the EUSDR is the second macro-regional strategy after the same action regarding Baltic Sea Region (European Union Strategy for the Danube Region). An action plan of the EUSDR put the accent on the environmentally sustainable way and takes into account the impacts of climate change at a basin scale . The Danube River Protection Convention (DRPC) represents a political framework for cooperation and transboundary water management for the entire river basins. Danube River Basin District Management Plan (DRBM plan), adopted in 2009, constitutes an ample analysis of the main pressures, especially human induced, in order to improve water quality at the basin level (ICPDR 2013). Disconnection of floodplains from the river is one of the main problems underlined by the DRBM plan. The important role of the wetlands ecosystems within a more complex biodiversity background is recognized, as is the floodplain retention function at the flood events. Reconnection and restorations of different areas have been identified to ensure biodiversity and some implementation steps were discussed at the basin scale.

The latest documentary information on hazard and flood risk is the Danube Atlas—Hazard and Risk Maps, a result of the cross-border Danube Flood-risk project (www.danube-floodrisk.eu). The atlas is the result of international cooperation between 8 nations with territories in the Danube River basin, including Austria, Italy, Slovakia, Hungary, Croatia, Serbia, Bulgaria and Romania . Flood hazard maps were produced for three scenarios: Floods with 30 % probability of exceedance (high probability), floods with 100 % probability of exceedance (medium probability) and floods with 0.01 % probability of exceedance (low probability). Regarding the possibility of increasing floods as consequence of climate change the project delivered an important basis for future concepts: The risk areas identified and mapped in the atlas show the problem area if climate change will increase the floods. Thus, the project is fundamental for developing long term strategies regarding the impact of climate change along the lower Danube.

The “Romanian Waters” National Administration is the public authority that manages the hydrologic infrastructure system through the work of the 11 water management units found in its administration. On the basis of hazard and risk maps, “Romanian Waters” National Administration will establish flood risk management plans coordinated at the level of the water management units, until December 2015, according to Directive 2007/60/EC also known as “Flood Directive”.

Romania recently completed a LIDAR -based Digital Elevation Model (DEM) for the entire Danube floodplain, including the delta . The high resolution DEM enables hydraulic scenarios to be envisioned for different flood levels. According to existing data, from the entire floodplain of 445,000 ha, three categories have been proposed as targets for the future: 43.3 % for agricultural fields; 40.8 % areas used as mixed (agricultural/polders and water storage ); 15.9 % areas for natural restoration (Nichersu 2009; Fig. 2). The conclusion is that 84.1 % of the lower Danube floodplain will remain agricultural. The mixed category of 40.8 % is quite ambiguous since it retains an agricultural function and stores water only during floods events. After the floods in 2006, despite new studies and warnings from the academic community and civil society (e.g., Vădineanu 2001; Iordache 2005; Stanciu et al. 2008), the management plan stipulates that embanked enclosures will be largely maintained. Restoration with an extensive rewilding program, however, will be the only economic solution. This is because, while embankments are useful for agriculture over a short period, a permanent loss of biodiversity is unacceptable to future generations.

13 Instead of Conclusions: Antipa—Environmentalist Avant-la-lettre

After more than a century, the strategy envisioned by Grigore Antipa (1895, 1907, 1910) still remains the only forward-looking solution for the economic exploitation of the Danube floodplain. The principle of rotating polders with alternation of agricultural crops and floodable areas, not only offers the best economic benefits, but simultaneously also allows for a better preservation of the fluvial ecosystem . Antipa warned that solutions for the management of Danube floodplain should not be imported from other regions of the world with different climatic conditions, hydrology or geomorphology . To copy these solutions and implement them on the lower course of the Danube would be “a great misfortune, which would make us lose the little we have”. The floodplain improvement system should be based on the local geomorphology of the floodplain, and the conservation of large and permanent lakes (Antipa 1910) is a feature that should be respected. Antipa’s caution is desirable in any environmental engineering or restoration approach and contrasts markedly with the dominant attitude of the communist regime and its will to control nature at any price. The radical transformation of the Danube floodplain affected not only the region, but also the mindset of the local population. The key to returning the floodplain to its natural state is in restoring the collective environmental memory of its people.