Navigation – Plan du site

The influence of river training on channel changes during the 20th century in the Lower Siret River (Romania)

L’influence des aménagements sur la métamorphose fluviale du Siret inférieur (Roumanie) au XXe siècle
Florence Salit, Gilles Arnaud-Fassetta, Liliana Zaharia, Malika Madelin et Gérard Beltrando
p. 175-188

Résumés

Cette étude s’attache à analyser la dynamique fluviale sur un siècle d’un cours d’eau fortement anthropisé, en lien avec les politiques d’aménagement et la variabilité climatique. Ce travail se concentre sur le secteur inférieur du Siret, avant dernier affluent du Danube (130 km de long et débit moyen annuel de 210 m3/s à la station de Lungoci). L’objectif de cet article est 1) de déterminer et d’analyser les modifications morphologiques du Siret inférieur du début du XXe siècle à 2010 et 2) d’en déterminer les principales causes. L’étude se fonde sur l’analyse de données historiques (textes et cartes), des données hydrologiques et climatiques et l’historique de l’aménagement de la rivière. À partir d’une analyse statistique et de la création d’un SIG, les évolutions hydro-morphologiques ont été quantifiées et la part de chacun des facteurs explicatifs (climat, occupation du sol, aménagements) discutés. Les principaux résultats montrent une contraction majeure de la bande active (- 46 % de 1940 à 2010) et une simplification du style fluvial particulièrement après 1970. Ces modifications sont expliquées avant tout par les stratégies d’aménagement de lutte contre les inondations (chenalisation, reboisement des berges et barrages) plutôt que par la variabilité climatique.

Haut de page

Texte intégral

This study formed part of a PhD thesis completed by Florence Salit. This research was supported by the EGIDE BRANCUSI – Campus France – program “Système hydrogéomorphologique et risques associés : impacts et aménagement territorial” (2011- 2012). We would like also to thank the anonymous reviewers for the helpful comments.

1Article soumis le 9 mai 2014, reçu sous sa forme révisée le 11 septembre 2014, accepté le 21 novembre 2014

1. Introduction

2Throughout Europe, many rivers have undergone widespread channel adjustments since the beginning of the 20th century, in particular incision and narrowing (Petts et al., 1989). The origin of these adjustments is widely discussed in the scientific community, due to the cumulative impact of both natural and human driven factors on fluvial system (Downs et al., 2013). Many authors (Bravard et al., 1997; Surian, 1999; Winterbottom, 2000; Liébault and Piégay, 2001, 2002; Arnaud-Fassetta, 2003; Swanson et al., 2011; Armaş et al., 2012; Ziliani and Surian, 2012; Siché and Arnaud-Fassetta, 2014) showed that, since the 19th century, human interventions such as dams, channelisation, in-stream mining, flow diversion, and land-use changes have affected the variables (water and sediment discharges; Schumm, 1977) controlling the hydromorphology of rivers. Each human intervention affects these variables in a different way and their effects may be cumulative but could also offset each other. The gravel mining led to a decrease of sediment whereas flow diversion involved a decrease of channel flow. Dams impose different changes of variables of control that depend on the decreasing peak runoff and sediment loads (Petts and Gurnell, 2005). These changes led to a complex response of channel adjustment. The decrease of sediment supply could lead to the narrowing and the incision of the riverbed and increased the disconnection between the channels and their alluvial plains (Gurnell, 1995). Some of authors (Liébault and Piégay, 2002; Wyżga, 2008) showed that there was no clear evidence that in the last decades, the climate could be regarded as the major factor of these morphological evolutions. Other authors (Pisŭt, 2002; Uribelarrea et al., 2003; Rădoane et al., 2010; Rădoane et al., 2013) have shown that climate change and associated changes in floods frequency and magnitude are the most important factor influencing morphological dynamics. Human interventions would only have accentuated these trends created by climate.

3The Romanian rivers have known the same trend. Both riverbed incision and channel narrowing were documented by several authors (Rădoane et al., 2010; Perşoiu and Rădoane, 2011; Armaş et al., 2012; Rădoane et al., 2013) but the origins of these adjustments are still widely discussed. In Romania, the context of engineering works in the riverbeds and floodplains is quite particular: an intense phase of river arrangement took place from the 1970s decade until the 1990s. Before this phase the engineering works (such as dyke) were very localised and specific. Today, the water management strategy goes through a transitional period between the needs of the economy (in-stream mining or water supply) and the implementation of the EU Water Framework Directive and Flood Directive. One of the aims of these directives is to implement an integrative management of water resource and flood at catchment scale (i.e. consider downstream effects on any riverbeds projects) and to preserve or/and restore good quality of water bodies.

4The aim of this paper is to determine and analyse the morphological changes of the Lower Siret River (Romania) in a diachronic way from the beginning of the 20th century to 2010, in order (i) to define the hydro-morphological changes of the river and (ii) to investigate the causes of these changes and especially the impact of engineering works. Our final goal is to assess the impact of the past management strategy on the Lower Siret River in the frame of a future implementation of a new EU River strategy.

2. Field context and engineering history

5The Siret river basin, located east of the Carpathians, has the largest catchment of the country with a total area of 42,900 km² and is one of the last tributaries of the Danube River. The mean annual flow is 210 m3/s at the Lungoci station (situated 88 km upstream of the Danube confluence). The largest part of the basin (48%) belongs to the Moldavian platform, made up of marls, sands, sandstones, gravel, and limestone (Rădoane et al., 2013). The basin morphology varies between a relatively deep valley upstream with a mean slope of 2%, and a plane morphology downstream, with a slope of less than 0.2% (Ministerul Mediului, 2009). The area covered with forests represents 37% of the basin surface (based upon Corine Land Cover data in 2006). The Siret catchment includes 291 dams and reservoirs built between the 1950s and 2007 for hydropower energy, irrigation, water supply, flood control, and fishing. Only seven reservoirs, designed to provide hydroelectric power and to flood control, are located in the Siret River. Just one dam (Movileni), achieved in 2007 is located on the Lower Siret River (fig. 1B).

6The study area is a 140-km long river reach located on the lower part of the catchment (fig. 1A). The elevation is between 60 m and 3 m a.s.l. The main tributaries are the Putna River, the Râmnicu Sărat River, the Buzău River and the Bârlad River. The Lower Siret River has a sandy bed with a slope of 0.02%. The majority of the land use is covered by arable lands whereas the forest cover represents almost 6% of the studied sector but only less than 5% of the total forest cover in the whole catchment.

Fig. 1 – Studied reaches in the Siret catchment.
Fig. 1 Localisation de la zone d’étude dans le bassin du Siret.

Fig. 1 – Studied reaches in the Siret catchment. Fig. 1 – Localisation de la zone d’étude dans le bassin du Siret.

A: Location of TR1 to TR4 reaches selected for the morphological study and engineering works. B: Study area; 1: river and lake; 2: dyke; 3: dam; 4: reaches; 5: sub-reaches; 6: hydrometric station.
A : Localisation des TR1 à TR4 sélectionnés pour l’analyse morphologique et localisation des aménagements. B : Zone d'étude ; 1 : rivière et lac ; 2 : digue ; 3 : barrage ; 4 : tronçon ; 5 : sous-tronçon ; 6 : station hydrométrique.

7The study area was divided into four reaches (fig. 1B) depending on local, hydrographical (tributaries) and anthropogenic (engineering works, bridges) context, and on available historic maps. The first reach (TR1) is the well documented with eight maps and satellite images from AD 1891 to 2010. Its course has the highest slope (0.07%) and its channel is braided. This reach can be divided into three sub-reaches: the upper one (TR1.1) extents from Movileni dam to the beginning of the levees area (6.5 km); the median one (TR1.2) is constrained by levees for flood control and bank protection structures (4.75 km); the lower one (TR1.3 – 5.25 km) goes from the Siret bridge to the end of the oldest map (1891). The second reach (TR2) is 28-km long and includes three major confluences, which two of them (Putna River and Bârlad River) were modified in the 1970s by deviation and regulation. The river dynamics is constrained by levees for flood control, which impose a channel width of about 200-500 m. TR2 is divided into two sub-reaches depending on tributaries. The third reach (TR3), 60-km long, is characterised by a river channel constrained for a long time (before 1940) by levees. It is divided into three sub-reaches depending on tributaries, and fluvial pattern is meandering. Downstream, the fourth reach (TR4 – 37.5 km) has the lowest slope (0.01%) and is divided into two sub-reaches constrained by levees.

8Engineering control works are various in the Lower Siret River and they are designed to control the river dynamics and to mitigate the fluvial risk linked to flood hazard and bank erosion. Water management strategy in the catchment stems from three phases (from i to iii), which implementation can be explained by both political (kind of government) and following the occurrence of major floods (tab. 1). (i) Between 1940 and the end of the 1960s, the floodplain was seen as a resource for the exploitation of the water and the land. (ii) In 1970, a major flood event occurred (a 10-year flood with a peak discharge of 3186 m3/s at Lungoci station), which led to change the river management strategy. The river was more a danger for both population and economy. Consequently, a complete and systematic arrangement was implemented with 900 km of dykes in the whole area, flow diversion, regulation, and an extensive afforestation policy. (iii) In July 2005, the Lower Siret River experienced a 100-year flood with a peak discharge of 4650 m3/s, highest value recorded at the Lungoci station (fig. 1B). This hydrological event was catastrophic, causing 23 deaths and billions Euros of damages (Ministerul Mediului şi Gospodăririi Apelor, 2006). Moreover this flood highlighted the defects of structural measures (several breaches appeared) and locally, network of dykes are severely discussed. Nowadays, the floodplain is considered as a complex area between natural heritage and danger. The Lower Siret River is a European protected area for birds and wetlands, but it is also a resource (water and sediment supply) for the new direction of the Romanian’s economy. The Lower Siret River has experienced a quite long history a human intervention and the aim of this work is to determine the consequences on its morphology.

Tab. 1 – Evolution of the management strategies from 1940 to 2012 in the Lower Siret River.
Tab. 1Évolution des stratégies d’aménagement de 1940 à 2012 sur le Siret inférieur.

Phases

Perception of the river

Strategy

Human interventions

Engineering works in the studied reach

1940-1969

Resource

Restricted

Sanitation drainage

15,000 ha

Channelisation

133 km

Flood

1970

1970-2000

Danger

Complete and systematic

Channelisation

900 km levees

Dams reservoirs

6 (upstream reach 1977-92)

Regulation

7 CC 24 G 2CD

Afforestation

+71 km²

Bank protection

20 km

Flood

2005

2005-2012

Danger

Integrated

Dam reservoir

1 (2007)

Resource

In-stream mining

1,384,644 m3/a allowed*

Natural Heritage

Banks protection

15 km + 5G

* estimation for 7 gravel-mining operations.

CC: channel cut off; G: groynes; CD: channel diversion.

Data derived from thirty different sources among which official reports, historical engineering works, literature, and field survey.
Données issues de trente différentes sources parmi lesquelles des rapports officiels, l’historique des travaux d’aménagement, la littérature et des observations de terrain.

3. Data and methods

9The evolution of the Lower Siret River during the 20th century was analysed on the basis of historical sources (text and maps), records of human intervention in the catchment, and hydrological data in order to assess planar changes of the River. The choice was made to study only the planar evolution as a first step to then analyse the morphological changes of the Siret River.

3.1. Analysis of cartographic sources

10In order to get comparable morphological data and analyse the Lower Siret River plan view dynamics, data were digitised from three cartographic materials, the only ones available for the entire reach (1940, 1981, and 2010). The use of historical sources such as maps (Piégay and Schumm, 2003; Hohensinner et al., 2004; Zanoni et al., 2008; Michalkova et al., 2011) are essential to understand the nature and magnitude of river channels modifications. Changes in the Lower Siret channel were accounted through diachronic analysis interpreting historic maps. Unfortunately no cartographic material was available in the timing of flood event. The planform morphology of the studied reach was documented by (i) 1:20,000 scale historic map sheets called Planuri Directoare de Tragere from 1891 to 1940; (ii) 1:25,000 scale Romanian topographic maps from 1981; (iii) and Google Earth images from 2009 and 2010. All digital data were captured in Quantum GIS environment projected into Google Mercator. We can estimate a potential error in results of +/- 10 m due to differences in quality of sources and digitalisation. In order to be more accurate in the timing of morphological changes, a set of topographic maps (1970, 1971, and 1990) and orthophotoplans (2005), available only for TR1 and TR2 were qualitatively used. The small scale (more than 1:50,000) of these maps did not allow a quantification of channel change. Furthermore, geomorphological indicators were selected to assess the spatial-temporal modifications of the channel. These indicators are the mean channel length, the braided index (average number of anabranches perpendicular to the bed axis; Ashmore, 1991), the sinuosity index (SI), and the mean active-channel width (channel + alluvial bars). The various channel morphologies were grouped into four general types indicated by the SI (Leopold et al., 1964): SI < 1.05, single-thread; 1.05 < SI < 1.25, sinuous; 1.25 < SI < 1.5, very sinuous; SI > 1.5, meandering. The active-channel width was automatically measured in 546 transects located along the Lower Siret River each 250 m. The evolution of the land-use cover of studied area was measured on the mapping materials and Corine Land Cover from 2006. Only riparian forest was measured in this analysis due to its relevant hydromorphological role in the evolution of channel planform (Gurnell, 1995). The GIS construction was useful to quantify both channel changes and forest cover evolution. Number and state of present engineering works were listed by several field surveys from 2009 to 2012.

3.2. Analysis of hydroclimatic data

11The hydroclimatic analysis was performed using two kinds of data. First, in order to define temporal tendencies of precipitation, the annual cumulative precipitation of six meteorological stations implemented in the Siret catchment was used (fig 1A). Produced by the National Administration of Meteorology (ANM), the data cover the period 1961-2010. These data are the only available in the Siret catchment with no gap for that period of time. Then, hydrological analysis derived from a data series allowed us to identify temporal tendencies. Flow records from the stream-gauging station in Lungoci were used (fig 1B). Data series included daily average water discharges from which mean annual water discharge was calculated (Qmean; 1953-2010), maximum annual water discharge (Qmax; 1951-2011), daily average suspended-sediment discharge (Qsmean; 1971-2011) from which mean annual suspended-sediment discharge was calculated, and maximum annual suspended-sediment discharge (Qsmax; 1971-2011). Time series were measured by ANAR (Administraţia Naţională Apele Române – “Romania Waters” National Administration Siret Water Branch) and obtained from ANAR and from GRDC (Global Runoff Data Center). After a critical analysis of data sets, temporal variations were analysed using three complementary statistical methods: (i) the non-parametric Mann-Kendall test (Kendall, 1975) was used to test the independence of the consecutive elements of an equally spaced time series. The null hypothesis is H0: the series is random, and its alternative is H1: there is a trend in the series. The significance level was measured by the probability p < 0.01. (ii) The break test of Pettitt (1979) was performed to detect changes in the mean of the time series. The change-point called break point provides here the year where the mean changes in time series. The Pettitt test is a non-parametric test. The null hypothesis is that there is no break point in the discharge series. The significance level chosen is p < 0.01. (iii) The Hubert’s segmentation procedure (Hubert et al., 1989) was applied to detect multiple breaks in the time series. The principle is to cut the series into n segments (n > 1) such that the calculated means of the neighbouring sub-series significantly differ. To limit the segmentation, the means of two contiguous segments must be different to the point of satisfying Scheffé’s test. Moreover if the phase described by the test is under five years, the result is not considered relevant. The procedure gives the timing of the breaks. The test was applied automatically with Khronostat 1.01.

4. Results

4.1. Changes in channel planform

12The results of the morphologic analysis (fig. 2) show two noticeable evolutions over time: an overall narrowing and a simplification of channel form.

Fig. 2 – Changes in channel planform in the Lower Siret River from 1891 to 2010.
Fig. 2 – Évolution de la morphologie en plan du Siret inférieur de 1891 à 2010.

Fig. 2 – Changes in channel planform in the Lower Siret River from 1891 to 2010. Fig. 2 – Évolution de la morphologie en plan du Siret inférieur de 1891 à 2010.

Active channel width variations for TR1 (A) and for TR2, TR3 and TR4 (B); (C) Geomorphological indicators variations on TR1, TR2, TR3 and TR4 (SI: sinuosity index; BI: braided index; L: channel length in km.
Variations de la largeur de la bande active sur le TR1 (A) et sur les TR2, TR3 et TR4 (B). C : Variations des indicateurs géomorphologiques sur les TR1, TR2, TR3 et TR4 (SI : indice de sinuosité ; BI : indice de tressage ; L : longueur du chenal en km.

13On the Lower Siret River, the mean of channel widths decreased from 670 m in 1940 to 300 m in 2010. The channel rapidly narrowed (7 m/a) between 1940 and 1981 but the narrowing fell (2.4 m/a) between 1981 and 2010. Eighty-one percent of the narrowing occurred between 1940 and 1981 for the all area. Figure 2 shows the main results for each studied reach. The first reach (TR1) recorded the most accentuated and complex changes (fig. 2A): the mean active-channel width increased (+ 395 m) from 1891 to 1940, in the first sub-reach (TR1.1) whereas the mean active-channel greatly decreased (- 1467 m) from 1940 to 2010. TR1.2 and TR1.3 recorded another trend: the decrease of the mean active-channel is slight from 1891 to 1940 (respectively - 130 and - 109 m) and more accentuated from 1940 to 2010 (respectively - 592 and - 1028 m). The mean active-channel width generally decreased from 1940 to 1981 in the TR2 and TR3 (fig. 2B) even if the decrease is more pronounced in TR2. The mean active channel width remains almost the same for the last reach (TR4). Finally from 1981 to 2010 the active channel width slightly decreased in TR2 and TR3.2 but slightly increased in TR3.1 and TR3.3 TR4.

14The simplification of the channel pattern (fig. 2C) is first recorded by the decrease of the braided index in the two first reaches from 1940 to 2010 (- 1.42 and - 0.27 respectively) and then by the decrease of channel length (-10 km) in TR4 due to six cut-off meanders from 1940 to 2010. The change was limited on the third reach (TR3); only the sinuosity index increased in 1981 from very sinuous to meandering. It appears that the two last reaches (TR3 and TR4) recorded the less change over time.

15In order to provide more details for the narrowing and simplification period, we use a set of available data existing only for the two first reaches (TR1 and TR2). The maps of 1970, 1971 and 1990 integrated into the GIS allowed a comparison of the channel but did not allow a quantitative analysis of channel narrowing (because of the scale and the low quality of the maps). Narrowing produced a significant change in the channel pattern, from braided to single-thread especially on the TR1. Channel narrowing began between 1940 and 1970 but the majority of the decrease took place between 1970 and 1981 (fig. 3). The TR1.1 shows the most complex evolution: an avulsion occurred before 1981 and the numbers of anabranches greatly decreased between 1970 and 1971. These evolutions might be explained by a major flood in 1970 (3186 m3/s at Lungoci station). The pattern evolution is accentuated from 1990 to 2010 by the completion of the Movileni dam in 2007, even if we can notice a channel widening between 2005 and 2010. The evolution of the TR2.1 follows the same trend even if the variations (simplification of the channel pattern, cut-off meander; narrowing) are less accentuated. To conclude, the active channel on the TR1 and TR2.1 narrowed and the channel pattern in TR1.1 and TR1.2 changed from braided to single-thread between 1940 and 1981, especially from 1970.

Fig. 3 – Channel narrowing of the Lower Siret River (TR1-TR2 reaches) from 1891 to 2010.
Fig. 3 – Contraction de la bande active du Siret inférieur (TR1 et TR2) de 1891 à 2010.

Fig. 3 – Channel narrowing of the Lower Siret River (TR1-TR2 reaches) from 1891 to 2010. Fig. 3 – Contraction de la bande active du Siret inférieur (TR1 et TR2) de 1891 à 2010.

4.2. Changes in discharge and suspended-sediment dynamics

16The Mann-Kendall test does not support the existence of significant change in time (p < 0.01) for mean annual discharge and maximum annual discharge (tab. 2). The probability to reject the null hypothesis when it is true is high (45%), thus there is no trend in the time series. Nevertheless, the results of Pettitt test at the significance level (p < 0.01) show that there is a significant break point in the year 1968. The Hubert’s segmentation was applied in order to confirm this result. There is no significant break point in the maximum annual discharge because the period found by the test is under five years long. But two dates seem to be especially significant in the mean annual discharge: 1968 and 1984. The results of Mann-Kendall tests for suspended-sediment discharges (Qs) denote that the null hypothesis is rejected, so the series have a trend. Two phases appear in the time series with a break point in 1984 for the mean annual sediment discharge. Nevertheless, the break point is different and more significant for the maximum annual sediment discharge, with a decrease of the mean of 80%. The hydrological analysis lead to the definition of three phases (fig. 4): (i) 1953-1968: low mean annual and maximum annual Q; (ii) 1968-1984: high mean annual and maximum annual discharges (Q and Qs); (iii) 1984-2010: high variability of Q, denoted by standard variation and strong decreases of Qs especially after 1992.

Tab. 2 – Results of the statistical tests on mean and maximum water (Qmean and Qmax) and suspended-sediment (Qs mean and Qs max) discharges of the Lower Siret River at Lungoci station.
Tab. 2 – Résultats des tests statistiques sur les débits liquides moyens et maximaux (Qmean et Qmax) et sur les débits d’alluvions en suspension (Qs mean et Qs max) à la station de Lungoci sur le Siret inférieur.

Variables

Time

Mann-Kendall test*

Pettitt test *

Break point

Hubert segmentation

Mean

Standard deviation

Qmean

1953-2010 (57 years)

45.25

12.1

1968

1953-1967

160.5

45.1

1968-1983

267.9

56.7

1984-2009

196.2

86.6

Qmax

1951-2011 (60 years)

48.19

5.43

1968

-

-

-

Qs mean

1971-2011 (40 years)

<0.01

<0.01

1985

1971-1983

454.3

132.5

1984-2011

123.6

159.25

Qs max

1971-2011 (40 years)

<0.01

<0.01

1991

1971-1991

15000

13326.2

1992-2011

2512.6

3600.8

* Probability of correctly rejecting the null hypothesis (H0) when it is truth. Significant level p=0. 01 in %.

Italic: no significant break point; bold: break point.
En italique : absence de point de rupture significatif ; en gras : point de rupture.

Data: ANAR-Siret Water Branch, GRDC.
Source des données : ANAR- Organisme du Siret, GRDC.

Fig. 4 – Water and sediment fluxes of the Siret River at the Lungoci hydrometric station.
Fig. 4 – Variation des débits liquides et d’alluvions en suspension à la station de Lungoci sur le Siret inférieur.

Fig. 4 – Water and sediment fluxes of the Siret River at the Lungoci hydrometric station. Fig. 4 – Variation des débits liquides et d’alluvions en suspension à la station de Lungoci sur le Siret inférieur.

Mean annual discharge (Qmean) from 1953 to 2010 (A); Maximum annual discharge (Qmax) from 1951 to 2011 (B); Mean annual suspended-sediment discharge (Qs mean) from 1971 to 2011 (C); Maximum annual suspended-sediment discharge (Qs max) from 1971 to 2011 (D).
Débits liquides moyens annuels (Qmean) de 1953 à 2010 (A) ; débits liquides maximaux annuels (Qmax) de 1951 à 2011 (B) ; débits moyens annuels d’alluvions en suspension (Qs mean) de 1971 à 2011 (C) ; débits maximaux annuels d’alluvions en suspension (Qs max) de 1971 à 2011 (D).

Data: ANAR-Siret Water Branch, GRDC.
Données : ANAR-Organisme du Siret, GRDC.

5. Cause of channel changes

17Channel geometry adjusts over time in response to changes that affect water and sediment discharges. The causes of control-variables changes are multiple and complex. Climatic changes since the end of Little Ice Age associated with the human interventions such as dams, flow regulation, channelisation and in-stream gravel mining are cited as the main causes of channel changes (Bravard et al., 1997; Winterbottom, 2000). Increase of the precipitation induces directly the increase of water discharge. Dams and reservoirs reduce the sediment load downstream and the peak flow discharge (Petts and Gurnell, 2005). Flow regulation and embankment can induce the narrowing of the active channel and the reduction of braided index (Surian, 1999). The extent of riparian forest tends to fix the channel and reduce the sediment supply. All these causes can be coincident (Gregory, 2006) and it is possible to assess the role of each cause involving the narrowing of the active channel.

5.1. The influence of the climatic factor

18Most of research on precipitation tendencies in Romania began in the 1960s. It is thus difficult to assess the part of climatic factor before 1960 due to the lack of data (Ştefan et al., 2004; Apostol et al., 2012; Ştefănescu et al., 2013). We only know some hydrological events (1897, 1902, 1914, 1930, 1942) occurring in the Siret River (Mustătea, 2005). These events could induce a widening of the active-channel and explain the trend of the TR1.1 between 1891 and 1940. Nevertheless, we can assess the precipitation tendencies in the surrounding region. In their studies in Romanian Plain, Stefan et al. (2004) have demonstrated that there was no trend in annual precipitation during the period 1931-1999. Some periods (1943-1952; 1958-1968; 1980-1999) were marked by a meteorological and hydrological drought. Zaharia and Beltrando (2009) analysed the data of 20 meteorological stations for the period 1962-2006 for the Carpatic Arc region. They demonstrated by statistical analysis that (i) there was no significant trend (at monthly and annual timescale) in the variability of the rainfall but an alternation of dry and wet years, and (ii) the period 1969-1972 was wet dominated. Ştefănescu et al. (2013) have studied the frequency of heavy (more than 50 mm) and extreme (+ 100 mm) rainfall in 24 h from 1980 and 2010 in 230 meteorological stations of Romania. They demonstrated that heavy and extreme rainfalls are more frequent and intense since 2000. In the present paper, six meteorological stations derived from data produced by the National Administration of Meteorology (ANM) were analysed at the Siret catchment scale in order to define a trend or cycles in the annual precipitation (tab. 3A). The Mann-Kendall test does not support the existence of significant change over time (p < 0.01) for annual precipitation. The probability to reject the null hypothesis when it is true is high (45%), thus there is no trend in the time series. The p-value of Pettitt test is too high to consider significant the various break points. Deviations from average were calculated in order to determine cycles in time series. A positive “anomaly” appeared for the whole six stations for the period 1969-1972. The variability of annual precipitation increased from 1980. Wet years (1984, 1991, and 2005) alternated with dry years (1986, 1990, 2000 and 2007) with no tendency evidence. The low discharge noticed in the 1950-1960s as well as the increase of the discharge in the 1970s can be explained by the precipitation variation. However, morphological changes, and particularly the active-channel narrowing, observed after 1970 cannot be explained by a climatic factor because these years are a relatively high-discharge period (fig. 4).

Tab. 3 – The role of climatic conditions and land-cover use. Main characteristics of meteorological stations in the Siret catchment and results of statistical tests; data ANM.
Tab. 3 – Le rôle des conditions climatiques et de l’occupation du sol. Principales caractéristiques des stations météorologiques sur le bassin versant du Siret et résultats des tests statistiques ; données ANM.

Tab. 3 – The role of climatic conditions and land-cover use. Main characteristics of meteorological stations in the Siret catchment and results of statistical tests; data ANM. Tab. 3 – Le rôle des conditions climatiques et de l’occupation du sol. Principales caractéristiques des stations météorologiques sur le bassin versant du Siret et résultats des tests statistiques ; données ANM.

A: Evolution of the forest in the first reach (TR1) of the Lower Siret River according to topographical maps; B: CLC 2006.
A : Evolution de la forêt sur le premier tronçon (TR1) du Siret inférieur selon les cartes topographiques ; B : Corine Land Cover 2006.

5.2. Role of land-cover use

19Land-use changes may induce high hydromorphological responses in catchments. Land-use changes led to a decrease (soils covered by forest) or an increase (soils not covered by vegetation) of water and sediment discharges (Leopold et al., 1964; Boix-Fayos et al., 2007). A decrease of riparian forest may have increased both runoff and sediment supply whereas an extension of the riparian forest and its stability in time can have induced the narrowing of the active channel (Kondolf et al., 2002). According to Giurescu (1976), in 1976, the forest covered just one third of what it was in the early 20th century in the country. This intense deforestation was accused of provoking major floods in the 1970s, that’s why the government conducted a strong afforestation policy (Zaharia, 1998) after 1970, even if this policy began to a lesser extent in the 1930s. We do not have the exact data for the entire Siret River catchment but we can assess, thanks to topographical maps (1891, 1940 and 1981) and Corine Land cover data (2006), the evolution of the forest cover on the Lower Siret River between 1940 and 2006, and especially on the first reach (TR1) between 1891 and 2006. The forest cover in the first reach is not representative of the cover in whole catchment, but the variations from the beginning of the century are similar in the Siret catchment. From 1940 to 1981, the forest cover doubled (+ 70 km²) in the Lower Siret floodplain. Table 3B shows the evolution of the forest evolution more particularly on the first reach (TR1). The forest cover doubled from 1940 to 1970 and increased (+ 78.2%) from 1971 to 1981 due to flood protection policy. This extension is most of the time an extent of the riparian forest made of poplars. Two hypotheses can be made: (i) land-use change in the 1970s explains the narrowing of the active channel and the simplification of channel pattern by fixing banks and avoiding divagation; (ii) afforestation policy is part of a more extensive program of channel and flow regulation. The forest expansion was synchronous with the active-channel narrowing. The second hypothesis seems to be the more accurate because of two reasons. First the channel geometry cannot adjust its morphology so quickly. Second, thanks to technical engineering reports of 1970’s we understand that afforestation policy was generally used to reduce the extension and the divagations of the riverbed.

5.3. Decisive impact of human intervention and arrangement

20Seven dams located in the Siret River were built between 1977 and 2007, and the last two (Călimaneşti and Movileni) date from 1992 and 2007. The decrease in suspended-sediment discharge series cannot be explained by the dam’s completion, which cannot be the first factor of the active-channel narrowing in the 1970s. First, narrowing phenomenon began earlier than dam’s construction. Nevertheless, dams have accentuated the decrease of suspended-sediment after the year 1991. Second, both the rivers Putna and Buzău are tributaries that brought the highest values of suspended-sediment discharge in the Siret River (Rădoane and Rădoane, 2005), and they are located downstream of all the dams. Consequently, the dams can just explain the break point in suspended-sediment discharge between 1984 and 1991.

21The dyke system appears as one of the most important factors impacting river morphology but it is variable, depending on reaches. The TR3 was channelised before 1940, thus there is no data to define a change due to engineering works. We assume that this early channelisation intended to protect the only foot bridge from the entire Lower Siret River upstream the confluence. After the major flood of 1970 and the highest annual maximum water discharges recorded from 1969 to 1975, the Romanian government decided to implement a strong flood protection strategy by building earth levees. That’s why from the 70s (most probably after 1974 – SGA Vrancea, 1974), the river was channelised in TR2 and TR4 in order to protect lands and villages from floods. These dykes appear to be one of the first control factors of the limitation of channel migration. The first reach (TR1) is the most affected by narrowing and simplification of channel pattern. Active-channel narrowing began before 1981 especially in two sectors (fig. 5A): (i) between 0-5 km, a possible avulsion and the abandonment of a right-bank branch caused a strong narrowing; (ii) the implementation of a dyke system upstream of the Siret Bridge constrained the channel. The active channel was stable and narrowed from 1981 upstream of the bridge (TR1.2) whereas a channel widening appears upstream and downstream this area (TR1.1 and TR1.3; fig. 5B).

Fig. 5 – Evolution of the active channel width in the TR1 reach of the Lower Siret River from 1940 to 2010 (A) and comparison of the active-channel morphological evolution from 1891 to 2010 (B). The complex arrangements of sub-reach TR1.2 upstream of the Siret bridge (C).
Fig. 5 – Évolution de la largeur de la bande active du Siret inférieur sur le TR1 de 1940 à 2010 (A) et comparaison de l’évolution morphologique de la bande active de 1891 à 2010 (B). Aménagement complexe du sous-tronçon TR1.2 en amont du pont du Siret (C).

Fig. 5 – Evolution of the active channel width in the TR1 reach of the Lower Siret River from 1940 to 2010 (A) and comparison of the active-channel morphological evolution from 1891 to 2010 (B). The complex arrangements of sub-reach TR1.2 upstream of the Siret bridge (C). Fig. 5 – Évolution de la largeur de la bande active du Siret inférieur sur le TR1 de 1940 à 2010 (A) et comparaison de l’évolution morphologique de la bande active de 1891 à 2010 (B). Aménagement complexe du sous-tronçon TR1.2 en amont du pont du Siret (C).

1: Siret active channel 1981; 2: Siret active channel 1940; 3: dyke 1981; 4: railway; 5: groynes; 6: buried groynes; 7: embankments; 8: concrete pylons and slabs.
1 : bande active du Siret en 1981; 2 : bande active du Siret en 1940 ; 3 : digue ; 4 : voie ferrée ; 5 : épis ; 6 : épis enterrés ; 7 : enrochement ; 8 : pylônes et plaques de béton armé.

22The second subreach (TR1.2) had known an extremely rapid channel contraction between 1970 and 1981 due to the building of engineering structures network (fig. 5C): twenty-two groynes buried or not and 4 km of rip-rap and concrete pylons were erected. These structures aim to avoid lateral erosion and channel widening in order to protect the bridge and arable lands near the villages. The groynes were made of 2x2 m concrete blocks perpendicular to the channel axis. Some of them were deeply destroyed and fell into the river due to lateral, bank erosion (fig. 6A). The single-thread channel shows evidence of erosion and widening, especially since 2005: in TR1 between 2005 and 2010, the mean width of the active-channel recorded an increase of 11% and a mean lateral erosion of 59 m (with a maximal value of 409 m recorded in TR1.1; Salit, 2013). This trend seems to be confirmed thanks to the field surveys between 2009 and 2012. The right-bank erosion was evaluated up to the number of fallen blocks (fig 6A). Downstream from the Siret bridge (TR1.3), the lateral erosion reached 2 m/a. During this period, the 2010 flood event and the increase of in-stream-mining (fig. 6B) may have accentuated both lateral erosion and widening. In sum two phenomena occurred on the first reach (TR1) from 1970 to nowadays due to human interventions: (i) from 1970 to 1981 the active channel width narrowed due to regulation and channelisation; (ii) since 2005 the active-channel width recorded a widening due to the combine effects of floods, dam and in-stream mining.

Fig. 6 – Evaluation of the lateral erosion in the Lower River Siret by quantification of destroyed groynes.
Fig. 6 – Évaluation de l’érosion latérale sur le Siret inférieur par le dénombrement des épis détruits.

Fig. 6 – Evaluation of the lateral erosion in the Lower River Siret by quantification of destroyed groynes. Fig. 6 – Évaluation de l’érosion latérale sur le Siret inférieur par le dénombrement des épis détruits.

A: Note the seven fallen blocks (14 m) pointed by *. B: In-stream mining in the active channel of the Lower Siret River.
A : Les sept blocs tombés (14 m) sont indiqués par *. B : Extraction de graviers en lit mineur sur le Siret inférieur.

Photograph: F. Salit, October 2011.
Photographies : F. Salit, octobre 2011

23On the Lower Siret River, human interventions such as dykes, dam and in-stream mining deeply modified the morphological pattern of the river. Climatic conditions and the associated hydrological events may have accentuated the morphological trend imposed by the overriding influence of the river engineering practices.

6. Discussion and conclusions

24The results from this analysis of morphologic changes in the Siret River over time show an overall narrowing and simplification of channel form, but the trend of variations is different between reaches (fig. 7A). The first two reaches and especially TR1 show the major evolution (braided channel to single thread; narrowing) from 1970 to 1981, whereas the morphological dynamics of TR3 and TR4 are less pronounced. Several factors could explain these differences. First the initial patterns of reaches were different (braided for the first two; meandering for the last two) and a braided river is more sensitive at short and medium term to a change of water and sediment discharges (Schumm, 1977; Sambrook Smith et al., 2006). Then the engineering works appears earlier in the TR3 that limits dynamics of the riverbed. The combined actions of various engineering works in the 1970s associated with afforestation policy and dam completion after 1990s could explain the evolution of the TR1 and TR2.The input of suspended load by main tributaries in TR3 and TR4 could explain the less morphological changes/dynamics of the two last reaches, but engineering works lead to a stabilisation of all parameters.

Fig. 7 – The compared trend of the active channel width on the different reaches of the Lower Siret River from 1940 to 2010.
Fig. 7 – Évolution comparée de l’évolution de la largeur de la bande active sur l’ensemble des tronçons du Siret inférieur de 1940 à 2010.

Fig. 7 – The compared trend of the active channel width on the different reaches of the Lower Siret River from 1940 to 2010. Fig. 7 – Évolution comparée de l’évolution de la largeur de la bande active sur l’ensemble des tronçons du Siret inférieur de 1940 à 2010.

A: Relative weight of the factors controlling the geomorphological evolution of the Siret active channel from the beginning of the 20th century to 2010. B: Upward arrow: increasing mean channel width; downward: decreasing mean channel width. 1: no data; 2: no weight or opposed factor; 3: no weight; 4: aggravating factor; 5: important factor; 6: major factor.
A : Poids relatif de chacun des facteurs contrôlant l’évolution morphologique de la bande active du Siret inférieur du début du XXe siècle à 2010. B : Flèche montante : augmentation de la largeur moyenne de la bande active ; Flèche descendante : diminution de la largeur moyenne de la bande active . 1 : pas de données ; 2 : pas de poids voire même facteur contraire ; 3 : pas de poids ; 4 : facteur aggravant ; 5 : facteur important ; 6 : facteur majeur.

25K.J. Gregory (2006) pointed out five questions dealing with the human role in changing river channels: what, how, when, where, and why. This paper answers some of these questions: in the Lower Siret River, morphological adjustments, especially active-channel narrowing, can be first of all explained by engineering interventions rather than just climatic variability, particularly in the 1970s. Engineering works were built in order to control the river dynamics and location and to protect lands and population from “the harmful effects of the water” (Hâncu, 1976). But there is no single answer to the spatial and temporal variations of the morphological channel adjustments in the Lower Siret River. The weight of each factor, climatic and human, in the active-channel dynamics is synthesised in fig. 7B. Five phases of change can be distinguished:

26(i) 1900-1950: The lack of climatic data before the 1960s in the Siret catchment permits only speculation. The engineering works really appear in 1940 but ramped up after 1970. Thus, only climatic drivers and land-use changes could lead to the widening of the bed during the first part of the century such as we noticed on the TR1.1. Thus, climatic conditions and deforestation could lead to active-channel widening by an increase of water and suspended-sediment discharges. From 1930, the beginning of afforestation policy could lead to reduce the weight of land use cover in the dynamics of active-channel width.

27(ii) 1950-1969: active-channel width increased less quickly due to various factors (decrease of precipitation, afforestation) and it is difficult to determine the most important one.

28(iii) 1970-1983: channel engineering combined with afforestation led to active-channel narrowing. The climatic factor (increase of precipitation) cannot explain this trend since higher discharges from increase rainfall should lead to an increase of channel capacity. The management strategy following the 1970 flood led toward a narrower, single-thread, less complex channel structure.

29(iv) 1984-1991: dams and channelisation contributed to the active-channel narrowing.

30(v) 1992-2010: channelisation led to the planform stability of the active channel, which can only adjust its morphology by incision. The in-stream mining developed since 2005 seems to exacerbate channel incision even if no data are available for the moment (fig. 6B). The 2005 "reference" flood event changed the positive vision of the river engineering strategy, especially in the Lower Siret River (Salit et al., 2013). The future engineering works (such as groynes) strive to protect the riverbanks from lateral erosion causing by all previous engineering works. This lateral erosion threatens villages and since 2009, seven new projects of banks protection were planned. However it seems that this dynamics lead to justify in-stream mining. Intense in-stream mining is motivated at a regional scale in in-stream mining environmental impact assessment (EIA) by several explanations: it aims avoiding lateral erosion and dredging river channel in order to increase its capacity and then prevent floods (Salit and Ioana-Toroimac, 2013). This strategy appears in contradiction with European directives and further investigations should be conducted to assess the morphological effects of in-stream mining in this area. The Lower Siret River experienced severe changes since the 1970s due to combined human factors. During this period, climatic variability and the associated hydrological events can be considered as triggering factors, which caused the activation of a systematic and complete water management strategy. In the Lower Siret catchment, management strategy plays a key role in the dynamics of the fluvial system more than any other variable. The major flood of 2005 involved a similar political will with beneficial effects that it will be possible to evaluate in the coming years.

Haut de page

Bibliographie

Apostol L., Machidon D., Machidon O., Buruiana D. (2012) Characteristics of precipitations trends in the area of Iaşi, Bacău and Suceava airports over the last 50 years. Aerul şi Apa: Componente ale Mediului, 252-258.

Armaş I., Nistoran D.E.G., Osaci-Costache G., Braşoveanu L. (2012) – Morpho-dynamic evolution patterns of Subcarpathian Prahova River (Romania). Catena 100, 83-99.

Arnaud-Fassetta G. (2003) – River channel changes in the Rhone Delta (France) since the end of the Little Ice Age: geomorphological adjustment to hydroclimatic change and natural resource management. Catena 51, 141-172.

Ashmore P.E. (1991) How do gravel bed rivers braid? Canadian Journal of Earth Sciences 28, 326-341.

Boix-Fayos C., Barbera C.G., Lopez-Bermudez F., Castillo V.M. (2007) – Effects of checkdams, reforestation and land-use changes on river channel morphology: case of study of the Rogavita catchment (Murcia, Spain). Geomorphology 91, 103-123.

Bravard J.-P., Amoros C., Pautou G., Bornette G., Bournaud M., Creuzé des Châtelliers M., Gibert J., Peiry J.-L., Perrin J.-F., Tachet H. (1997) River incision in south-east France: morphological phenomena and ecological effects. Regulated Rivers: Research & Management 13, 75-90.

Downs P.W., Dusterhoff S.R., Sears W.A. (2013) Reach scale channel sensitivity to multiple human activities and natural events: Lower Santa Clara River, California, USA. Geomorphology 189, 121-134.

Giurescu C.C. (1976) Istoria pădurii româneşti. Ed Ceres, Bucureşti, 392 p.

Gregory K.J. (2006) The human role in changing river channels. Geomorphology 79, 3-4, 172-191.

Gurnell A.M. (1995) Vegetation along river corridors: hydrogeomorphological interactions. In Gurnell A.M., Petts G.E. (Eds.): Changing River channels. Wiley, Chichester, 237-260.

Hohensinner S., Habersack H., Jungwirth M., Zauner G. (2004) Reconstruction of the characteristics of a natural alluvial river floodplain system and hydromorphological changes following human modifications: the Danube River (1812-1991). River Research and Applications 20, 25-41.

Hâncu S. (1976) – Regularizarea albiilor rîurilor. Ceres, Bucureşti, 143 p.

Hubert P., Carbonnel J.-P., Chaouche A. (1989) Segmentation des séries hydrométéorologiques. Applications à des séries de précipitations et de débits de l’Afrique de l’Ouest. Journal of Hydrolology 10, 349-367.

Kendall M.G. (1975) Rank correlation methods. Charles Griffin, London, 202 p.

Kondolf G.M., Piégay H., Landon N. (2002) – Channel response to increased and decreased bedload supply from land use change: contrasts between two catchments. Geomorphology 45, 35-51.

Leopold L.B., Wolman M.G., Miller J.P. (1964) Fluvial processes in geomorphology. Freeman, San Francisco, 522 p.

Liébault F, Piégay H. (2001) – Assesment of channel changes due to long-term bedload supply decrease, Roubion River, France. Geomorphology 36, 167-186.

Liébault F., Piégay H. (2002) Causes of 20th century channel narrowing in mountain and piedmont rivers of southeastern France. Earth Surface Processes and Landforms 27, 425-444.

Michalkova M., Piégay H., Kondolf G.M., Greco S. (2011) Longitudinal and temporal evolution of the Sacramento River between Red Bluff and Colusa, California (1942-1999). Earth Surface Processes and Landforms 36, 257-272.

Ministerul Mediului şi Gospodăririi Apelor (2006) – Raport privind efectele inundaţiilor şi 21 fenomenelor meteorologice periculoase produse în anul 2005. Bucureşti, 40 p.

Ministerul Mediului (2009) – Contribuţii la dezvoltarea strategiei de management al riscului la Inundaţii, raport: Studiu privind inundaţiile, schemă directoare şi plan de investiţii pentru Siretul inferior, Versiune finală – Partea I. 253 p.

Mustăţea A. (2005) Viituri excepţionale pe teritoriul României. Geneză şi efecte. Editura Institutului Naţional de Hidrologie şi Gospodărire a Apelor, Bucureşti, 280 p.

Perşoiu I., Rădoane M. (2011) Spatial and temporal controls on historical channel responses - Study of an atypical case: Someşu Mic River, Romania. Earth Surface Processes and Landforms 36, 1391-1409.

Pettitt A.N. (1979) – A non-parametric approach to the change-point problem. Journal of Applied Statistical Science 28, 126-135.

Petts G.E., Muller H., Roux A.-L. (Eds) (1989) Historical change of large alluvial rivers: Western Europe. Wiley, Chichester, 355 p.

Petts G.E., Gurnell A.M. (2005) Dams and geomorphology: Research progress and future directions. Geomorphology 71, 27-47.

Piégay H., Schumm S.A. (2003) – System approaches in fluvial geomorphology. In Kondolf G.M., Piégay H. (Eds.) Tools in Fluvial Geomorphology. Wiley, Chichester, 105-134.

Pisŭt P. (2002) – Channel evolution of the pre-channelised Danube River in Bratislava, Slovakia (1712–1886). Earth Surface Processes and Landforms 27, 369-390.

Rădoane M., Rădoane N. (2005) Evoluţia actuală a piemontului pericarpatic moldovenesc. Analele Universităţii “Stefan cel Mare” Suceava. Anul XIV, 11-19.

Rădoane M., Pandi G., Rădoane N. (2010) Contemporary bed elevation changes from the Eastern Carpathians. Carpathian Journal of Earth and Environmental Sciences 5, 49-60.

Rădoane M., Obreja F., Cristea I., Mihailă D. (2013) – Changes in the channel-bed level of the eastern Carpathian rivers: Climatic vs. Human control over the last 50 years. Geomorphology 193, 91-111.

Salit F. (2013) De l’eau, des digues et des hommes. Approche géographique du risque inondation sur le Siret inférieur (Roumanie). Thèse de doctorat en géographie, université Paris-Diderot (Paris 7), 360 p.

Salit F., Beltrando G., Zaharia L. (2013) The impact of the July 2005 flood to river management: study case of the Lower Siret River (Romania). In Arnaud-Fassetta G., Masson E., Reynard E. (Eds.): European Continental Hydrosystems under Changing Water Policy. Friedrich Pfeil Verlag, München, 273-279.

Salit F., Ioana-Toroimac G. (2013) – Actual in-stream mining in alluvial rivers: geomorphological impact and European legislation. Proceedings of the 3rd International Geography Symposium (GEOMED), Antalya-Kemer-Turkey, June 10-13 2013. Inkilap, Istanbul, 201-210.

Sambrook Smith G.H., Best J.L., Bristow C.S., Petts G.E. (Eds.) (2006) – Braided Rivers: Process, Deposits, Ecology and Management. Special Publication 36 of the IAS, Wiley-Blackwell, Oxford, 396 p.

Siché I., Arnaud-Fassetta G. (2014) – Anthropogenic actions since the end of the Little Ice Age: A critical factor controlling the last ‘fluvial metamorphosis’ of the Isonzo River (Italy, Slovenia). Méditerranée, 122, 183-200.

Schumm S.A. (1977) The fluvial system. Wiley, New York, 338 p.

SGA Vrancea (1974) Lucrări de apărare împotriva inundaţiilor de pe rîurile Siret, Putna, Rm. Sărat şi Leica.

Stefan S., Ghioca M., Rimbu N., Boroneant C. (2004) Study of meteorological and hydrological drought in southern Romania from observational data. International journal of climatology 24, 7, 871-881.

Ştefănescu V., Ştefan S., Georgescu F. (2013) Spatial distribution of heavy precipitation events in Romania between 1980 and 2009. Meteorological Applications 21, 3, 684-694.

Surian N. (1999) Channel changes due to river regulation: the case of the Piave River, Italy. Earth Surface Processes and Landforms 24, 1135-1151.

Surian N., Rinaldi M. (2003) – Morphological response to river engineering and management in alluvial channels in Italy. Geomorphology 50, 307-326.

Swanson B.J., Meyer G.A., Coonrod J.E. (2011) Historical channel narrowing along the Rio Grande near Albuquerque, New Mexico in response to peak discharge reductions and engineering: magnitude and uncertainty of change from air photo measurements. Earth Surface Processes and Landforms 36, 885-900.

Uribelarrea D., Pérez-Gonzáles A., Benito G. (2003) Channel changes in the Jarama and Tagus rivers (central Spain) over the past 500 years. Quaternary Science Reviews 22, 2209-2221.

Winterbottom S.J. (2000) Medium and short-term channel planform changes on the Rivers Tay and Tummel, Scotland. Geomorphology 34, 195-208.

Wyżga B. (2008) A review on channel incision in the Polish Carpathian rivers during the 20th century. In Habersack H., Piégay H., Rinaldi M. (Eds.): Gravel-bed Rivers VI: From Process Understanding to River Restoration. Elsevier, Amsterdam, 525-556.

Zaharia L. (1998) – Tendances dans l’évolution des transferts de matières en suspension dans les Subcarpates de Courbures en relation avec les modifications du milieu. Géomorphologie : Relief, Processus, Environnement, 1, 3-15.

Zaharia L., Beltrando G. (2009) Variabilité et tendances de la pluviométrie et des débits de crue dans la région de la courbure de l’arc Carpatique (Roumanie). Geographia Technica, Numéro Spécial, 471-476.

Zanoni L., Gurnell A., Drake N., Surian N. (2008) Island dynamics in a braided river from analysis of historical maps and air photographs. River Research and Applications 24, 1141-1159.

Ziliani L., Surian N. (2012) Evolutionary trajectory of channel morphology and controlling factors in a large gravel-bed river. Geomorphology 173-174, 104-117.

Haut de page

Annexe

Version française abrégée

La description et l’analyse des changements historiques dans les systèmes fluviaux anthropisés ont mobilisé de nombreux chercheurs. Une tendance générale à l’incision des rivières a été observée, aboutissant à un rétrécissement du lit depuis 150 ans dans de nombreuses rivières européennes (Liébault et Piégay, 2002 ; Surian et Rinaldi, 2003 ; Wyżga, 2008). Un débat est né de ce constat entre les tenants d’une origine anthropique de ces changements et les tenants d’une origine plus naturelle. L’objectif de cet article est de déterminer et d’analyser 1) les dynamiques hydrologiques et morphologiques décelables sur le Siret inférieur (Roumanie) du début du XXsiècle à 2010 et 2) quels facteurs peuvent intervenir pour expliquer ces variations.

La première partie de l’article présente le site d’étude du Siret inférieur (fig. 1) ainsi que le contexte particulier de l’aménagement des cours d’eau en Roumanie. Le bassin du Siret, le plus grand de Roumanie, est situé à l’est des Carpates. Avec un débit moyen de 210 m3/s à la station de Lungoci, le Siret est l’avant dernier affluent majeur du Danube. Le Siret a fait l’objet d’un aménagement intensif (900 km de digues) particulièrement après l’inondation majeure de 1970 (tab. 1).

La seconde partie de l’article s’attache à décrire les données et méthodes utilisées, afin de définir les variations hydrologiques et morphologiques du Siret inférieur. Les données hydrologiques utilisées sont des séries chronologiques de débit – liquide et matières en suspension (MES) – à la station de Lungoci. Un test de tendance (Mann-Kendall) et deux tests de rupture ont été appliqués pour détecter une singularité unique (Pettitt) ou multiple (Hubert) à toutes les séries de données. L’analyse diachronique des dynamiques en plan du Siret a été réalisée à partir des données spatiales de 1940, 1981 et 2010, digitalisées, intégrées et analysées dans un SIG. Le secteur a été découpé en quatre tronçons homogènes et dix sous-tronçons (fig. 1). Quatre indices ont été calculés : la longueur du chenal, la largeur de la bande active, l’indice de tressage et l’indice de sinuosité.

La troisième partie présente les différents résultats des analyses hydrologique et morphologique. On constate une diminution de 46 % de la largeur de la bande active de 1940 à 2010 sur l’ensemble du secteur (fig. 2). La majorité (81 %) de cette réduction se déroule entre 1940 et 1981. Ces tendances masquent les différentes évolutions au sein du secteur. Associés à une contraction majeure de la bande active, les deux premiers tronçons connaissent une simplification de leur style fluvial, qui passe d’un style en tresses à rectiligne, surtout à partir de 1970 (fig. 3). Trois phases ont été identifiées dans l’analyse hydrologique (fig. 4) : 1) 1951/53-1968 : des débits faibles et donc des crues peu nombreuses ; 2) 1968-1984 : des débits liquides et d’alluvions en suspension moyens et maximaux importants ; 3) 1985-2010/11 : forte variabilité des débits liquides. Les crues sont moins fréquentes que lors de la période précédente mais plus intenses. Les débits de MES connaissent une baisse majeure surtout à partir de 1991.

La quatrième partie met en avant les causes de ces différentes évolutions constatées. Plusieurs facteurs explicatifs sont analysés : les facteurs climatiques, les changements d’occupation du sol, plus particulièrement l’évolution des surfaces boisées, et les aménagements dont les barrages et les digues. Chacun de ces facteurs peut avoir une action combinée dans le système fluvial (Gregory, 2006) ou même avoir des effets qui s’annulent. L’analyse des cumuls de précipitations montre que les débits plus faibles des années 50 et 60 tout comme la hausse des débits constatée dans les années 70 jusqu’au milieu des années 80 s’expliquent en partie par le régime des précipitations. Ainsi, les premiers résultats obtenus suggèrent que le rétrécissement de la bande active constaté dans les années 70-80 ne peut être lié à une période de crues moins intenses ou moins fréquentes puisque cette période correspond à une hausse moyenne des débits en lien avec une augmentation des précipitations. Les phases de variations des débits liquides tout comme la période de contraction de la bande active ne correspondent pas aux années de mise en place des barrages (1977-1992) ; ces derniers ne semblent être que des facteurs aggravants du phénomène. La chenalisation et plus particulièrement les travaux de régularisation effectués sur le Siret apparaissent comme l’une des causes principales de la contraction de la bande active. Plusieurs éléments en témoignent : 1) la concomitance entre période de contraction de la bande active et période d’aménagement intensif et 2) la réduction brusque en un ou deux ans de la largeur de la bande active, notamment pour les TR1.2 et TR1.3 (fig. 5).

Les variations spatiales (entre les tronçons) et temporelles (sur environ un siècle) des ajustements du chenal ne permettent pas d’apporter une réponse unique aux différentes évolutions constatées. La figure 7 présente la synthèse des causes possibles et du poids de chacune dans l’évolution de la largeur de la bande active du Siret inférieur de 1900 à 2010, avec une attention particulière accordée à la période 1970-1980. Plusieurs périodes sont distinguées : 1) 1900-1950 : le climat et le couvert forestier sont considérés comme des facteurs explicatifs majeurs dans l’élargissement de la bande active car aucun aménagement n’est présent lors de cette période ; 2) 1950-1969 : le ralentissement de l’élargissement de la bande active à partir des années 50 n’est qu’une hypothèse : la baisse des cumuls pluviométriques, qui se traduisent par une baisse des débits moyens annuels, peut conduire au ralentissement de l’élargissement ; 3) 1970-1983 : la chenalisation massive combinée aux mesures intensives de reboisement des corridors fluviaux favorisent une contraction de la bande active et un accroissement des vitesses d’écoulement. Les conditions climatiques (forts cumuls pluviométriques annuels conduisant à de nombreuses crues) n’ont pas pu jouer un rôle dans cette tendance. Au contraire, elles auraient plutôt eu un rôle inverse en favorisant la mise en place des politiques d’aménagement. Mis en service à la fin de cette période seulement, les barrages n’ont fait qu’accentuer cette tendance ; 4) 1984-1991 : les barrages associés au maintien de l’endiguement ont accentué la réduction de la bande active ; 5) 1992-2010 : la chenalisation a favorisé la stabilité de la bande active. Les conditions climatiques, même si elles sont très variables (événements hydro-climatiques intenses sur la période), ne peuvent jouer sur ce marqueur, le lit étant contraint dans un réseau de digues.

Le Siret inférieur est un exemple de rivière fortement anthropisée dont les caractéristiques morphologiques actuelles sont très influencées par la vague d’aménagements des années 70. Les politiques passées de gestion de la rivière ont modifié le comportement du système fluvial et, en ce sens, les nouvelles stratégies de gestion des inondations doivent intégrer ces nouvelles caractéristiques hydrogéomorphologiques pour une meilleure appréhension des interactions Homme-Environnement.

Haut de page

Table des illustrations

Titre Fig. 1 – Studied reaches in the Siret catchment. Fig. 1 Localisation de la zone d’étude dans le bassin du Siret.
Légende A: Location of TR1 to TR4 reaches selected for the morphological study and engineering works. B: Study area; 1: river and lake; 2: dyke; 3: dam; 4: reaches; 5: sub-reaches; 6: hydrometric station. A : Localisation des TR1 à TR4 sélectionnés pour l’analyse morphologique et localisation des aménagements. B : Zone d'étude ; 1 : rivière et lac ; 2 : digue ; 3 : barrage ; 4 : tronçon ; 5 : sous-tronçon ; 6 : station hydrométrique.
URL http://geomorphologie.revues.org/docannexe/image/11002/img-1.jpg
Fichier image/jpeg, 928k
Titre Fig. 2 – Changes in channel planform in the Lower Siret River from 1891 to 2010. Fig. 2 – Évolution de la morphologie en plan du Siret inférieur de 1891 à 2010.
Légende Active channel width variations for TR1 (A) and for TR2, TR3 and TR4 (B); (C) Geomorphological indicators variations on TR1, TR2, TR3 and TR4 (SI: sinuosity index; BI: braided index; L: channel length in km.Variations de la largeur de la bande active sur le TR1 (A) et sur les TR2, TR3 et TR4 (B). C : Variations des indicateurs géomorphologiques sur les TR1, TR2, TR3 et TR4 (SI : indice de sinuosité ; BI : indice de tressage ; L : longueur du chenal en km.
URL http://geomorphologie.revues.org/docannexe/image/11002/img-2.jpg
Fichier image/jpeg, 648k
Titre Fig. 3 – Channel narrowing of the Lower Siret River (TR1-TR2 reaches) from 1891 to 2010. Fig. 3 – Contraction de la bande active du Siret inférieur (TR1 et TR2) de 1891 à 2010.
URL http://geomorphologie.revues.org/docannexe/image/11002/img-3.jpg
Fichier image/jpeg, 1,3M
Titre Fig. 4 – Water and sediment fluxes of the Siret River at the Lungoci hydrometric station. Fig. 4 – Variation des débits liquides et d’alluvions en suspension à la station de Lungoci sur le Siret inférieur.
Légende Mean annual discharge (Qmean) from 1953 to 2010 (A); Maximum annual discharge (Qmax) from 1951 to 2011 (B); Mean annual suspended-sediment discharge (Qs mean) from 1971 to 2011 (C); Maximum annual suspended-sediment discharge (Qs max) from 1971 to 2011 (D). Débits liquides moyens annuels (Qmean) de 1953 à 2010 (A) ; débits liquides maximaux annuels (Qmax) de 1951 à 2011 (B) ; débits moyens annuels d’alluvions en suspension (Qs mean) de 1971 à 2011 (C) ; débits maximaux annuels d’alluvions en suspension (Qs max) de 1971 à 2011 (D).
URL http://geomorphologie.revues.org/docannexe/image/11002/img-4.jpg
Fichier image/jpeg, 864k
Titre Tab. 3 – The role of climatic conditions and land-cover use. Main characteristics of meteorological stations in the Siret catchment and results of statistical tests; data ANM. Tab. 3 – Le rôle des conditions climatiques et de l’occupation du sol. Principales caractéristiques des stations météorologiques sur le bassin versant du Siret et résultats des tests statistiques ; données ANM.
Légende A: Evolution of the forest in the first reach (TR1) of the Lower Siret River according to topographical maps; B: CLC 2006.A : Evolution de la forêt sur le premier tronçon (TR1) du Siret inférieur selon les cartes topographiques ; B : Corine Land Cover 2006.
URL http://geomorphologie.revues.org/docannexe/image/11002/img-5.png
Fichier image/png, 76k
Titre Fig. 5 – Evolution of the active channel width in the TR1 reach of the Lower Siret River from 1940 to 2010 (A) and comparison of the active-channel morphological evolution from 1891 to 2010 (B). The complex arrangements of sub-reach TR1.2 upstream of the Siret bridge (C). Fig. 5 – Évolution de la largeur de la bande active du Siret inférieur sur le TR1 de 1940 à 2010 (A) et comparaison de l’évolution morphologique de la bande active de 1891 à 2010 (B). Aménagement complexe du sous-tronçon TR1.2 en amont du pont du Siret (C).
Légende 1: Siret active channel 1981; 2: Siret active channel 1940; 3: dyke 1981; 4: railway; 5: groynes; 6: buried groynes; 7: embankments; 8: concrete pylons and slabs. 1 : bande active du Siret en 1981; 2 : bande active du Siret en 1940 ; 3 : digue ; 4 : voie ferrée ; 5 : épis ; 6 : épis enterrés ; 7 : enrochement ; 8 : pylônes et plaques de béton armé.
URL http://geomorphologie.revues.org/docannexe/image/11002/img-6.jpg
Fichier image/jpeg, 760k
Titre Fig. 6 – Evaluation of the lateral erosion in the Lower River Siret by quantification of destroyed groynes. Fig. 6 – Évaluation de l’érosion latérale sur le Siret inférieur par le dénombrement des épis détruits.
Légende A: Note the seven fallen blocks (14 m) pointed by *. B: In-stream mining in the active channel of the Lower Siret River. A : Les sept blocs tombés (14 m) sont indiqués par *. B : Extraction de graviers en lit mineur sur le Siret inférieur.
URL http://geomorphologie.revues.org/docannexe/image/11002/img-7.jpg
Fichier image/jpeg, 1,9M
Titre Fig. 7 – The compared trend of the active channel width on the different reaches of the Lower Siret River from 1940 to 2010. Fig. 7 – Évolution comparée de l’évolution de la largeur de la bande active sur l’ensemble des tronçons du Siret inférieur de 1940 à 2010.
Légende A: Relative weight of the factors controlling the geomorphological evolution of the Siret active channel from the beginning of the 20th century to 2010. B: Upward arrow: increasing mean channel width; downward: decreasing mean channel width. 1: no data; 2: no weight or opposed factor; 3: no weight; 4: aggravating factor; 5: important factor; 6: major factor. A : Poids relatif de chacun des facteurs contrôlant l’évolution morphologique de la bande active du Siret inférieur du début du XXe siècle à 2010. B : Flèche montante : augmentation de la largeur moyenne de la bande active ; Flèche descendante : diminution de la largeur moyenne de la bande active . 1 : pas de données ; 2 : pas de poids voire même facteur contraire ; 3 : pas de poids ; 4 : facteur aggravant ; 5 : facteur important ; 6 : facteur majeur.
URL http://geomorphologie.revues.org/docannexe/image/11002/img-8.jpg
Fichier image/jpeg, 982k
Haut de page

Pour citer cet article

Référence papier

Florence Salit, Gilles Arnaud-Fassetta, Liliana Zaharia, Malika Madelin et Gérard Beltrando, « The influence of river training on channel changes during the 20th century in the Lower Siret River (Romania) », Géomorphologie : relief, processus, environnement, vol. 21 – n° 2 | 2015, 175-188.

Référence électronique

Florence Salit, Gilles Arnaud-Fassetta, Liliana Zaharia, Malika Madelin et Gérard Beltrando, « The influence of river training on channel changes during the 20th century in the Lower Siret River (Romania) », Géomorphologie : relief, processus, environnement [En ligne], vol. 21 – n° 2 | 2015, mis en ligne le 01 janvier 2016, consulté le 26 juin 2017. URL : http://geomorphologie.revues.org/11002 ; DOI : 10.4000/geomorphologie.11002

Haut de page

Auteurs

Florence Salit

Université de Cergy-Pontoise UMR 8586 du CNRS (PRODIG) 75205 Paris cedex 13 France (florence.salit@u-cergy.fr). Tél. : 33 (0)1 34 25 64 09 ; Fax : 33 (0)1 44 07 75 63.

Articles du même auteur

Gilles Arnaud-Fassetta

Université Paris-Diderot-Sorbonne-Paris-Cité – UMR 8586 du CNRS (PRODIG) – 75205 Paris cedex 13 – France (gilles.arnaud-fassetta@univ-paris-diderot.fr).

Articles du même auteur

Liliana Zaharia

Faculty of Geography University of Bucharest Bd Bălcescu N° 1, Secteur 1 010041 Bucharest Romania (zaharialili@hotmail.com).

Articles du même auteur

Malika Madelin

Université Paris-Diderot-Sorbonne-Paris-Cité UMR 8586 du CNRS (PRODIG) 75205 Paris cedex 13 France (malika.madelin@univ-paris-diderot.fr).

Gérard Beltrando

Université Paris-Diderot-Sorbonne-Paris-Cité UMR 8586 du CNRS (PRODIG) 75205 Paris cedex 13 France (beltrando@univ-paris-diderot.fr).

Articles du même auteur

Haut de page

Droits d’auteur

© Groupe français de géomorphologie

Haut de page
  • Logo CNRS - Institut des sciences humaines et sociales
  • Logo Groupe français de géomorphologie
  • Revues.org