The guest editor would like to thank Nicole Limondin-Lozouet, Pascale Ruffaldi and the editorial board of Quaternaire for having accepted and supported this special issue. All the reviewers are also greatly acknowledged for their constructive comments which allowed improvement of the quality of the papers.

1-Introduction

1Fluvial systems are strongly influenced by environmental changes, both of natural and anthropogenic origin (Vandenberghe et al., 2010). Their present-day operation should thus be considered within a continuum with their history and evolution, from relatively recent anthropogenically forced changes back to the creation of the modern valley sand to predecessor drainage patterns, where these can be determined (Cordier & Bridgland, 2012). Fluvial archives of various kinds provide the evidence from which the environmental and hydro-climatic history of this past evolution can been reconstructed. These archives first include the landforms that characterize fluvial landscape and typically result from a long-term predominant erosional trend (Bridgland et al., 2009). These features are strongly coupled with fluvial sediments, the variety and characteristics of which reflect the past dynamics of river systems. Within the sedimentary bodies that constitute a large proportion of fluvial archives, archaeological artefacts can be found and organic materials can be preserved in various ways (charcoal, pollen and wood, together with faunal and microfaunal remains), providing evidence of human activity and of environmental conditions (White et al., 2013; Bridgland & White, 2014). In addition to these natural records, it is worth mentioning historical archives: early maps, recorded descriptions, flood marks, etc. The latter should be considered in a different way since they are often not observable in the field and because they can be of discontinuous character and possibly inaccurate. Nonetheless, they can provide complementary information about past river evolution.

2The importance of natural fluvial archives lies in what they can reveal about the past, at various timescales. They are, in essence, disconnected from the active fluvial system, even if this does not exclude, for example, post-depositional reworking of valley-floor and low-terrace deposits by floods and/or changes of river course. Put differently, these archives have not been affected by any fluvial process since their formation, albeit that slope processes, weathering and/or pedogenesis (including cryogenic activity) have commonly affected them (especially older sediments). Such fluvial archives provide a palaeoenvironmental record that extends back into the historical period, the Holocene interglacial, the Pleistocene and even, in some fluvial systems, a large part of the Cenozoic (e.g. Schoorl & Veldkamp, 2003; Pastre, 2005; Maddy et al., 2007). As is commonly emphasized, they facilitate understanding of the present-day functioning of fluvial systems (Herget et al., 2007), and will be of great value in the task of anticipating the potential response of these systems to future global change (Gregory et al., 2006).

3The study of fluvial palaeoenvironments is the principal theme for several international research initiatives (Herget et al., 2007), amongst which are included the FLuvial Archives Group (FLAG), the GLObal Continental PalaeoHydrology (GLOCOPH) group, and the Past Global Changes (PAGES) project-Focus 4, which includes the former working group Land-Use and Climate Impacts on Fluvial Systems (LUCIFS). FLAG was created as a thematic working group of the Quaternary Research Association in 1996, with the aim of bringing together researchers interested in the study of past fluvial activity and how this reveals responses to various types of extrinsic forcing (Cordier et al., 2014a). GLOCOPH was founded in 1991 as a subcommission of the International Union for Quaternary Research (INQUA), and is dedicated to the study of various aspects of the hydrological cycle in the past, in relation to rivers, lakes, and groundwater systems, at all temporal time scales but with special emphasis on Late Pleistocene and Holocene change. GLOCOPH is re-defining its aims within the research of past fluvial processes and extreme hydrological events, envisaging continuation of its fruitful collaboration with FLAG within the focus area of “Palaeohydrology and Fluvial Archives”, which remains under the aegis of the Terrestrial processes, Deposits and History commission (TERPRO) of INQUA.

4The PAGES project was initiated in 1991; amongst its various research groups, LUCIFS was dedicated, prior to the 2012-2013 Focus 4 (Human-Climate-Ecosystem Interactions), to the study of catchment-scale fluvial activity -including sediment fluxes- since the onset of agriculture, and the responses shown to human influence and climate change. In addition to the output from these groups, it is worth noting the research dedicated to fluvial geoarchaeology, in part undertaken within the framework of the Geoarchaeology Working Group of the International Association of Geomorphology, although IGCP 449 and IGCP 518 (International Geoscience programme projects from the first decade of the 21^st Century) incorporated fluvial geoarchaeology working groups (Bridgland et al., 2006, 2007; Mishra et al., 2007; Westaway et al., 2009).

5The aim of the present paper is to explore the details and rationale of the approaches conducted within the different research structures, which can be considered representative of the diversity of palaeoenvironmental data from fluvial archives. It will also include a synthesis of the contributions to the present special issue, which are a reflection of this diversity.

2 - The value of various types of fluvial archives

6Within the framework of FLAG research, the morpho-stratigraphical study of fluvial landforms is of considerable importance. Unravelling the formation and evolution of river valleys, which are predominantly erosional landforms, is a primary theme for FLAG. Valleys typically exhibit river terraces on their flanks, in variable patterns of distribution and preservation, and these record stages in the progressive deepening of the valley (Bridgland & Westaway, 2008a,b). There are several elements, both depositional and morphological, shared by most fluvial terraces. Commonly there is an important sedimentary component, which can be a repository of palaeoenvironmental information or of plant and animal fossils that can provide indications of age or other dating evidences (see below).

7In terms of morphology, the upper surfaces of terrace landforms, readily visible in the landscape, facilitate the recognition and downstream correlation of terraces. The basal contacts between terrace sediments and underlying bedrock can also be regarded as morphological features, albeit buried, and can provide important information about past fluvial regimes: e.g., the contrast between widespread flat bedrock platforms, often underlying braided-river gravel floodplains, and constricted-width channel incisions, the latter arising from single-channel rivers and often representing warmer climatic periods. The modern surfaces of terraces are likely to have been affected by runoff erosion, the accumulation of non-fluvial sediments (e.g., slope deposits and loess), or anthropogenic modification (e.g., for agriculture, road building or quarrying), all of which may have disturbed and/or disguised the original topography. In contrast the bedrock surface contact is much less likely to have been affected by erosion and can be more a more reliable measure of morpho-stratigraphy, as it corresponds with the limit of the fluvial incision and the beginning of accumulation. It can thus be of considerable value for the reconstruction of valley evolution, although its form will only be known in detail where there has been extensive borehole investigation or quarrying. In subsiding areas, or at the floodplain scale, the landscape reveals less about fluvial history, although there may be geomorphological evidence for minor erosional periods; one or more generations of palaeochannels are commonly preserved, providing information about the past drainage regime and pattern. Such landforms are, for example, well developed in the Lower Rhine (Busschers et al., 2007) and subsiding structural basins in the Upper Rhine (Lauer et al., 2010), and in the Great Hungarian Plain (Gábris & Nádor, 2007), which was visited during the FLAG 2008 meeting.

8Fluvial landforms are less central to the research priorities of GLOCOPH or PAGES-Focus 4. However, as the reference study area for the latter is the fluvial catchment (including not only the valley floor but also the slopes), geomorphology within the catchment is a fundamental component of background context for the study of past changes.

9Whatever the topic, fluvial sediments represent a principal source of palaeoenvironmental information, providing data of use in the three main hydro-geomorphological approaches to the reconstruction of past river systems: palaeohydrography, palaeohydrology and palaeohydraulics (Arnaud-Fassetta et al., 2010). Hydrography primarily refers to fluvial style, which can be interpreted from sedimentary facies and is of considerable relevance to there search undertaken within FLAG, as it makes it possible to assess the influence of climate: e.g., contrasts between wide, sheet-like spreads of terrace gravel that represent braided gravel-bed rivers and constricted channel-fill deposits, the latter arising from single-channel rivers, often during warmer climatic periods. The composition of fluvial sediments may also provide evidence for catchment and drainage-pattern changes, such as river-capture events or other diversions, as illustrated by examples from the Paris basin (Harmand & Cordier, 2012) and central-eastern England (White et al., 2010). In a more geoachaeological approach, the recognition of fluvial facies (as specific from marine or lacustrine) illustrates the role of fluvial systems in landscape evolution, in relation to human occupation (Ghilardi et al., 2011). From a hydrological view point, the particle size of alluvial terraces has been used (especially within the framework of GLOCOPH) to assess palaeodischarge during floods (Stokes et al., 2012b). Similarly, the study of palaeochannel size and form allowed Arnaud-Fassetta et al. (2010) to reconstruct fluvial hydraulic function, with reference, for example, to channel capacity. The same authors demonstrated that the three approaches hydrography, hydrology and hydraulics could usefully be combined to provide a better understanding of the historical evolution of river systems and their relations with past human societies, for example in terms of adaptation to fluvial-system changes or hazard management.

10Beyond their intrinsic characteristics, which relate directly to channel function, fluvial sediments can provide evidence for palaeoclimatic conditions, in particular with the recognition of cryoturbation features and pedogenically modified layers. Moreover, recent improvements in geochronological techniques have allowed direct dating of fluvial deposits (Wallinga, 2002; Rittenour, 2008; Cordier, 2010; Stokes et al., 2012b). Optically Stimulated Luminescence (OSL) dating, which can be applied to quartz and feldspars grains, has, during the last decade, become a common tool for dating sediments from the second half of the Middle Pleistocene (e.g. Schulte et al., 2008; Lauer et al., 2010; Briant et al., 2012; Harding et al., 2012; Cordier et al., 2014b) to the historical period (e.g. Cordier et al., 2012b; Medialdea et al., 2014). The dating of older sediments (back to the Pliocene) has been made possible by the development of cosmogenic nuclide dating, which has been successfully applied in various areas (Rixhon et al., 2011; Stange et al., 2013). An alternative method is the use of the Electron Spin Resonance (ESR) technique, which has been applied to several fluvial systems, particularly in Northern France and the surrounding areas (Antoine et al., 2007; Harmand et al., 2015). It is worth noting that, in addition to the direct dating of fluvial sediments, minimum or maximum ages can be derived from the dating of stratigraphically related deposits, such as volcanic layers (datable using K-Ar or Ar-Ar methodology, e.g. Pastre, 2005; Maddy et al., 2007; Veldkamp et al., 2012) or the U/Th dating of speleothems found in caves filled with fluvial sediments (e.g. Losson & Quinif, 2001) or of post-depositional calcareous (calcrete) cementation of fluvial deposits, typical of warmer-climate areas (e.g. Candy et al., 2004; Bridgland et al., 2012). Furthermore, dated speleothems from complex cave systems have been used to gauge the timing of progressive valley-floor lowering by fluvial erosion, based on the relation between speleothem formation and groundwater tables (Westaway, 2009, 2010). Such numerical ages are of fundamental importance in the study of past fluvial environments, since they are essential for assessing the role of various drivers such as climate or human societies (see below), as well as the rates of change. They can also be used, for example, to estimate rates of fluvial incision, providing new insight on the relation between fluvial systems and tectonic activity or epeirogenic uplift (Brocard et al., 2003; Cordier et al., 2006; Fuchs et al., 2014; Westaway & Bridgland, 2014). Beyond this, numerical ages represent a major tool to develop regional or global comparison between rivers, allowing the recognition of general trends in their evolution, but also specific behaviours that can be linked either to intrinsic forcing, or to local conditions (Vandenberghe, 2008; Cordier et al., 2014a).

11Fluvial sediments are also of great interest as they are host to various types of palaeoenvironmental evidence. The main types are (1) palaeobotany, incorporating pollen, plant macrofossils (including wood fragments), macro- and micro-charcoal as well as phytoliths (Garnier et al., 2015), (2) fauna and microfauna, especially mammals (Antoine et al., 2000; Schreve & Bridgland, 2004; Bridgland et al., 2012), molluscs (Keen, 1990; Preece, 1999; Limondin-Lozouet & Antoine, 2006; White et al., 2013), insects (Coope, 2001, 2006) and ostracods (Griffiths, 2001; Bridgland et al., 2012; White et al., 2013), (3) archaeological artefacts, from prehistoric stone tools (Bridgland & White, 2014) to the remains of historical buildings and structures. Such evidence can also be used for dating, for example using radiocarbon or dendrochronology. Indeed, research performed under the aegis of GLOCOPH have, over a number of years, compiled radiocarbon databases from which have been identified periods of enhanced fluvial activity during the Holocene (Macklin et al., 2006). Palaeobotanical and faunal remains also provide complementary information about the environmental conditions at the time of deposition, which can be combined with numerical ages and palaeohydrological data (e.g., in relation to fluvial regime) to provide an overview of past fluvial landscapes. Research related to LUCIFS and PAGES-Focus 4 has also demonstrated that accumulated organic sediments on valley floors can provide an excellent record of the onset of agriculture, with the transition from natural flora (typically woodland in Western Europe) to a mosaic of wooded and open-landscapes formed as a result of forest clearance for farming purposes (Houben et al., 2006; Lespez et al., 2008; Le Jeune et al., 2012). The paper by Ferrio et al. (2015) provides an excellent illustration of the potential of combining the study of fluvial systems, vegetation, and chronology. It also underlines the importance of multidisciplinary research, bringing together geomorphologists/geologists and geochronologists, but also palynologists, palaeontologists/malacologists, and archaeologists.

12The study of archaeological remains from fluvial sediments has considerable potential. First, these remains have a chronological significance which, although not as precise as numerical dating, can be used both at the Pleistocene (glacial-interglacial) scale (Bridgland & White, 2014) and for the historical period (Cordier et al., 2012b). Second, archaeological remains can provide evidence for in-situ human occupation, opening the way to various geoarchaeological studies, with the aim of unravelling the relations between past societies and their environmental conditions (e.g. Ghilardi & Boraik,, 2011; Heyvaert et al., 2012; Arnaud-Fassetta et al., 2014; Carozza et al., 2014; Castanet et al., 2014; Howard et al., 2014).

13The combined study of valley-landform geomorphology, fluvial sediments, and the organic or archaeological remains preserved in the latter, improves our understanding of past fluvial systems over long-term (Pleistocene glacial-interglacial) and short-term (sub-Holocene-historical) timescales. This information is of value in understanding the role of external drivers as an influence on fluvial dynamics.

3 - Deciphering fluvial responses to various extrinsic forcing factors

14Tectonic forcing or, more precisely, forcing by crustal movements (not all of which are of tectonic origin; see for example Panin et al., 2015) is a key focus for research undertaken within the framework of FLAG and focusing on the Late Cenozoic, Pleistocene or even post-Pleistocene evolution of fluvial systems (Bridgland et al., 2009, 2010). The influence of eustasy is also an important consideration for fluvial reaches located in lowland areas (Wolf et al., 2014). On the other hand, crustal movements and eustasy are of little relevance within research related to GLOCOPH or PAGES-Focus 4, or for geoarchaeological studies (except in relation to sites located in river mouths or offshore (Coles, 1998; Bridgland et al., 1999).

15In contrast, interest in the fluvial response to climate change is common to most of the research devoted to past river systems. This fact should first be explained by the complexity of potential climatic forcing (Vandenberghe, 2003, 2008); the influence of climate is not only direct, on the basis that precipitation, like rivers, is part of the hydrological cycle, but also indirect, through the intermediaries of temperature (which controls evaporation, ice formation and melting, snow melting, formation of permafrost…), vegetation (which directly influences erosion and sediment supply on the slopes and within the rivers), human presence, etc. There is an important linkage between climatic forcing and crustal movements, through the waxing and waning of ice-sheet cover (controlled by climate) and the isostatic effect of this on crustal movements (Bridgland & Westaway, 2008b; Westaway, 2012). It is worth noting that the role of climate can be seen at various timescales, from the Pleistocene Milankovitch cycles (e.g. Nádor et al., 2003; Maddy et al., 2005; Bridgland, 2006) to Holocene Rapid Climate Change (e.g. Thorndycraft & Benito, 2006b; Macklin et al., 2010; Richardson et al., 2013) and to the flash-flood events occurring after heavy rainfall. The importance of climate is also strengthened by the multiple issues raised by 21^st century anthropogenic climate change, which is likely to be a strong influence (amongst many) on fluvial systems, with consequences for human societies in terms of droughts, floods etc.

16However, the meaning of “climate forcing” significantly varies according to the research framework within which it is applied. Research performed within FLAG typically refer to the glacial-interglacial cycles (Marine Isotope Stages) that characterize the Pleistocene, although there is also considerable attention paid to forcing at shorter timescales, generally spanning the Lateglacial and Holocene (e.g. Kasse et al., 2010; Wolf et al., 2014). It has been widely demonstrated that such climate cycles were responsible, especially in uplifted areas of the mid (temperate) latitudes, for the alternation of periods of accumulation and erosion, leading to the formation of the terrace staircases that now characterize the valley landscape (Bridgland et al., 2009). In the vicinity of the glaciated areas (e.g. northern Europe and central England), the changes in fluvial courses induced by the variable extent of the glaciers also led to a complex pattern of reorganization within the drainage network that has been extensively studied over many decades (e.g., Marks, 2004;White et al., 2010).

17Research on palaeohydrology, especially under the aegis of GLOCOPH, has also been applied to Pleistocene terraces (Stokes et al., 2012a). However, the focus has typically been on the Holocene (Benito & O’Connor, 2013); study of fluvial sediments preserved in the modern floodplain is based on the recognition of facies that are interpreted in terms of fluvial activity (from active channel deposits to the filling of palaeochannels by organic sediment). The presence of extensive organic material allows significant radiocarbon databases to be constructed and used to identify periods of high fluvial activity (coupled with a higher frequency of extreme events, which is the heart of the HEX project; Benito et al., 2013). Comparison between various countries makes it possible to identify similarities that may be interpreted as fluvial responses to global climatic forcing, related, for example, to Rapid Climate Change events (e.g. Starkel et al., 2006; Thorndycraft & Benito, 2006a; Macklin et al., 2010). Research on this topic has also been conducted as part of LUCIFS/PAGES-Focus 4, which also focuses on the Holocene, but largely concentrates on sediment fluxes (Dotterweich, 2008; Brown et al., 2009; Hoffmann et al., 2010). The clear aim is to isolate the climate signal from anthropogenic influences related to deforestation and the development of agricultural practices. Such research therefore often combines the study (and dating) of fluvial sediments with the analysis of slope deposits and has thus been of great interest in demonstrating the non-linear response of sediment transfer from the interfluves to the valley floor and then to the river, at centennial to millennial timescales, and the importance of sediment storage (e.g. Lang & Hönscheidt, 1999; Brown, 2009; Fuchs, 2012).This key result is consistent with fluvial-system research that has demonstrated the role of threshold effects and the potential inertia of such systems (Herget et al., 2007). Geoarchaeological studies within fluvial contexts also generally combine attention on climate and on human societies (Brown, 1997). The main difference is that they are strongly focused on local-scale studies of archaeological sites. Two main kinds of research may be distinguished, the first oriented towards palaeoenvironmental reconstruction, with an approach that is very similar to that proposed by PAGES-Focus 4. Even if based on the study of in-situ records that may not reflect the reality and diversity of the large fluvial systems in which these belong (Vis et al., 2010; Le Jeune et al., 2012; Pastre et al., 2014), this approach allows general trends of human influence to be recognized. The second main geoarchaeological approach is directly related to site context, with research aiming to unravel the local fluvial dynamics (e.g. channel migration) and the relations with human societies in terms of amenities, adaptation, and fluvial hazard (Cordier et al., 2012b; Franc & Vérot-Bourrély, 2014).

18This short review underlines the various approaches that can be applied to the study of fluvial archives in terms of research objectives and key drivers. This variety is well reflected in the diversity of papers included in this special issue.

4 - Content of the special issue

19This special issue includes contributions focusing on various fluvial systems in Europe (Russia, Germany, France, Spain) and Africa. Various timescales are considered, ranging from the Pleistocene to the Lateglacial-Holocene and recent decades.

20The first paper, by D. Harmand et al., focuses on the Pleistocene evolution of the Rivers Moselle and Meurthe in the Paris Basin and the Rhenish Massif (France, Germany). It is mainly based on the use of the Electron Spin Resonance (ESR) dating method applied to fluvial sediments of Lower to Middle Pleistocene age. This study provides complementary geochronological data additional to those already obtained from the youngest terraces of these systems (Cordier et al., 2014a). It also allows a renewed comparison with the neighbouring Meuse system, in particular concerning the formation of the Main Terrace Complex (Rixhon et al., 2011).

21The second paper, by A. Panin et al., leads us to the upper Dnieper valley in Western Russia. The authors combine geomorphology and geochronology to unravel post-LGM (Last Glacial Maximum) valley evolution in a 150 km long reach of the Dnieper. They provide evidence for a complex pattern of accumulation and incision, largely driven by the dynamics in the vicinity of the glaciated area and by climate. In particular, they show that the formation and subsidence of a glacial forebulge was responsible for deformation of the valley, with profound effects (albeit temporary in part) on the palaeohydrology. The authors also provide a new insight into the formation of residual hills of glacigenic material (dating from MIS 6) and of the large palaeochannels preserved along the valley. The latter are proved to be of Early Holocene age, i.e. younger than the large palaeochannels recognized in Central European rivers.

22The third paper, by A. Garnier et al., focuses on the palaeoenvironmental evolution along the Yamé River (Mali) between 3,000 and 1,000 BP. The authors associate morphological and sedimentological investigations, together with radiocarbon dating, to unravel the evolution of the valley floor and the slopes of this system. This allows them to identify a synchronous fluvial incision event that took place throughout the study area between 2,350 and 1,700 cal. ka BP, as a response to climate forcing. At other times, different parts of the fluvial system have responded in different ways, as shown by the recognition of various fluvial patterns. The authors finally provide new insights into the interactions between the fluvial system and the human settlements in the area during the second part of the Holocene.

23The final paper, by J.P. Ferrio et al., is focused on a very short timescale, corresponding to the last few decades. The authors combine tree ring analysis (dendrohydrology) with the study of carbon and oxygen isotopes preserved in the tree rings, in order to try to date and quantify past flood events by comparison with the isotopic precipitation signal. The paper presents the application of this method to a flash flood event that took place in central Spain in 1999. The authors show that past floods associated with high rainfall can be recorded in the isotope composition of tree rings and can be isolated from the response to average climate variables, opening the way to improved reconstructions of past flood events.

5 - Conclusion: linking the research structures

24The existence of several research groups devoted to the study of fluvial archives is primarily a historical artefact, each group having been formed on the basis of the definition of one or several specific topic(s) (including the pertinent related timescales, methods and objectives of study). All of these research structures manage to obtain significant visibility in the scientific community, by their affiliation to major research structure or projects (GLOCOPH, Pages-Focus 4, Working Group of Geoarchaeology, etc), and above all by the involvement in scientific events such as meetings/sessions or special issues in international audience journals.

25Due to the methodological improvements (especially in geochronology and modelling) and to the development of multidisciplinary projects, more and more studies fall outside the central brief of any particular research group. Thus many bridges have, especially in recent decades, been built between these groups. The present special issue illustrates this well, as it follows a session organized jointly by FLAG and GLOCOPH during the IAG Meeting in Paris, 2013, where geoarchaeologists were also present. The relations between FLAG and geoarchaeology are significant, especially at the Pleistocene timescale: the presence of artefacts within fluvial sediments gave rise to interdisciplinary research since the XIX^th century, through the combined work of Jacques Boucher de Perthes (1788-1868), Joseph Prestwich (1812-1896), John Lubbock (later Lord Avebury; 1834-1913) and l’abbé Henri Breuil (1877-1961), who pioneered Pleistocene fluvial geoarchaeology in the valley of the River Somme, northern France (see review in Bridgland, 2014); the need to interpret Lower and Middle Palaeolithic stone artefacts within the context of fluvial-system dynamics is a continued catalyst for interdisciplinary research (e.g., Antoine et al., 2010; Bridgland & White, 2014). Similar approaches have also been developed at the Lateglacial to Holocene timescale. In the timeframe of the last few millennia, the research performed within FLAG is very close to that conducted by geoarchaeologists and as part of PAGES-Focus 4. The main difference is that the former are based on the study of (archaeological) sites to obtain results that are extrapolated to the river catchment, while the latter aims to reconstruct fluvial activity directly at the catchment scale, with analyses focusing not only on valley-floor sedimentation, but also on the interfluves.

26This combination of specific analyses and complementary research is a guiding principle in establishing links between research structures and, beyond these, between the scientists involved in the study of fluvial archives, ensuring that cooperation continues, enabling a better understanding of the past processes and activities of fluvial systems globally, which will be invaluable as a means of predicting their immediate future evolution.