Characterization and quantification of Holocene colluvial and alluvial sediments in the Valdaine Region (southern France)

Abstract

This study presents a Holocene sediment budget for the Valdaine Region, located at the edge of the southern French Pre-Alps. Holocene colluvial and alluvial deposition are quantified based on existing and new field data. Average sediment thickness values were calculated for different landscape units, and available chronological data were used to make a time-differentiated sediment budget. Total Holocene colluvial deposition in the Valdaine (334 km²) amounts 167 × 10⁶ m³, while alluvial deposition in the Roubion and Jabron catchments (in total 610 km², including their catchment upstream the Valdaine) amounts 177 × 10⁶ m³. Especially, colluvial deposition is high (0.75 × 10⁶ g/m²) compared with other catchments. Three major deposition periods (8500–2000 bc, 700 bc–ad 800 and ad 1200–2000) and two periods of relative hillslope stability with local fluvial incision (2000–700 bc and ad 800–1200) set the framework for a time-differentiated sediment budget. Results show that depositional phases relate to intense land use and hillslope stability and fluvial incision to land abandonment. Catchment averaged colluvial deposition increases from 13 × 10⁻⁶ m³/m²/yr for 8500–2000 bc to 355 × 10⁻⁶ m³/m²/yr for ad 1200–2000, while alluvial deposition increases from 16 to 147 × 10⁻⁶ m³/m²/yr between the same time periods. A relationship with climatic fluctuations was not found because of the limited temporal resolution of the sediment budget. The derived sediment cascade model shows how alluvial sediments change from fine (silt and clay) to coarse (sand and gravel) after ad 1200. This went along with the establishment of a braided river pattern, indicating that the main source of sediment shifted to the mountainous headwaters. Further research including fingerprinting and modeling would be necessary to further understand the sediment budget and to more accurately quantify the different source areas and the export from the catchment.

Keywords

alluvium climate colluvium France geoarchaeology geomorphology Holocene land use sediment budget

Introduction

Over a Holocene timeframe, soil erosion and sediment deposition are important processes in many European catchments (Dotterweich, 2008; Notebaert and Verstraeten, 2010). The processes are important for shaping the Earth surface and transport sediments and related biochemical constituents from the upland toward the oceans, with intermediate sinks which act as buffers. An increasing number of studies have shown the importance of quantifying the different processes involved over different spatial and temporal scales, in order to understand the sediment cascade (e.g. Hoffmann et al., 2007; Notebaert et al., 2011a; Trimble, 1999, 2009; Verstraeten et al., 2009b). Under ‘sediment cascade’, we understand the transport of sediment through a catchment’s landscape, including sources, redistribution pathways, and sinks. Quantification of such processes is also important to assess the role of soil erosion and sediment burial in the global carbon cycle (e.g. Hoffmann et al., 2009, 2013; Van Oost et al., 2012).

Often sediment budgets are used to quantify the sediment cascade. To quantify, the sediment budget methodology can be used to give an accounting of the sources and sinks of sediment flows for a spatial unit over a given time period (Reid and Dunne, 2005; Slaymaker, 2003). Over recent years, an increasing number of sediment budgets spanning timescales of decades to millennia have become available (e.g. Hinderer, 2012). Such long-term sediment budgets allow to understand how sediment cascades change in response to environmental changes. In Europe, the majority of sediment budgets refers to timescales of years to decades, while sediment budgets referring to timescales of millennia deal with catchment in Germany, Belgium, and the Netherlands (e.g. De Moor and Verstraeten, 2008; Förster and Wunderlich, 2009; Fuchs et al., 2011; Hoffmann et al., 2007; Notebaert et al., 2009; Seidel and Mackel, 2007; Stolz, 2011; Verstraeten et al., 2009b). Among these studies, chosen budget elements vary and may range from sole quantifications of total floodplain storage (e.g. Stolz, 2011) to more detailed studies including different sink types (e.g. Houben, 2008; Macaire et al., 2002) and complex time-differentiated sediment budgets (e.g. Notebaert et al., 2011a, 2013; Trimble, 2009; Verstraeten et al., 2009b). For most catchments, sediment budgets show the large influence of anthropogenic land use changes on the sediment cascade (e.g. Trimble, 1999; Verstraeten et al., 2009a).

The results of these studies illustrate the non-linear nature of sediment flows between different components of the sediment cascade and how this leads to variable sedimentation rates over space and time (e.g. Fuchs et al., 2011; Notebaert et al., 2011a).

In contrast, only very few sediment budget studies are available for the landscapes of southern Europe (e.g. Dusar et al., 2011), although these provide classical examples of past to present human impacts on used landscapes. The French Valdaine region is an example of a south European landscape, and consequently, a wealth of archaeological and geo-archaeological data are available. This includes coring and profile data on Holocene colluvia, alluvia, and soil profile change, often with age information derived from archaeological artifacts or radiocarbon dating. So far, research has focused on until now site-specific reconstructions (e.g. Berger, 1996; Berger et al., 2002; Vital et al., 1999), particular cultural period synthesis (Berger and Brochier, 2006; Berger et al., 2007a), and relationships between geomorphology and archaeological site preservation (Berger, 2011; Verhagen and Berger, 2001). Quantification of sediment sinks and flows and a sediment cascade model are still missing.

The main objective of this paper is to quantify the Holocene sediment cascade for the Valdaine region by the construction of a time-differentiated sediment budget. By doing so, we aim at closing the gap between traditional sediment budget studies and site-specific geo-archaeological studies. This will be done by using geo-archaeological data to construct a sediment budget, and linking the sediment budget results to palaeo-environmental data in order to understand the temporal and quantitative variations in sediment redistribution.

Study area

This study focuses on the Valdaine region (334 km²), which forms the lower part of the Jabron and Roubion catchments (total 610 km²), situated west of the city of Montélimar, France (Figure 1). The Jabron river enters into the Roubion river near Montélimar, shortly upstream of its confluence with the Rhône river. The headwaters of both rivers are located in the mountainous calcareous Pre-Alps, while the Valdaine landscape lies in a lower situated structural depression with gently rolling hills.

Figure 1.

Study area. (a) Overview of the study area, with indication of the extent of the Valdaine region, the upper Roubion and Jabron catchments, and the selected tributary catchments (see text). The background colors represent the altitude. (b) Landscape units of the Valdaine region and indication of coring and profile data used in this study. A color version is available online.

Altitudes in the Valdaine region range between 70 m and c. 850 m a.s.l. The central part of the catchment has a less expressed relief, with altitudes between 70 m and c. 200 m a.s.l., and slopes angles are generally less than 2%. Catchment margins rise up to altitudes of 500–600 m, and even 850 m in the east. The slopes angles are often higher than 10%. Some limestone hills also stick out in the central Valdaine plain as ridges, with slopes > 10% and height differences of more than 50 m.

The region has a transitional climate between the Mediterranean, continental, and Atlantic climates (Blanchet, 1990). Mean annual precipitation amounts to 900–1000 mm (varying annually from 300 to 1700 mm; Blanchet, 1990; Hirsch and Vincent, 1966). There is a high inter-annual variability in precipitation pattern, with some years having large amount of precipitation in spring and fall (September to November), typically for a Mediterranean climate, and other years with precipitation more evenly spread over the year.

An extended dataset on archaeological sites exists in the Valdaine region, mainly originating from field walking surveys and rescue archaeology projects (such as Archaeomedes, ATP CNRS, TGV-Méditerranée). Settlement density was reconstructed, taking into account the erosion and reworking of archaeological deposits. Results show an intensive occupation history since the Neolithic Period (Berger, 1996, 2011). Settlement density peaks first during the middle Neolithic Period (c. 4500–3500 bc) (Berger, 1996). Malacological data indicate a forested landscape during the Early–Mid Bronze Age (c. 2300–1300 bc), followed by a progressive opening of the environment during the Late Bronze age (c. 1300–750 bc), while there was a general open landscape during the Iron Age and Roman Period (750 bc–ad 500) (Berger et al., 2007b). Settlement density data published in Berger (1996) and Berger et al. (2007b) show an alternation of less and more dense settlement patterns during the Bronze Age and Iron Age, while a peak in settlement density occurs during the High Roman Empire (ad 1–150). This is followed by a decrease in settlement density during the Low Roman Empire (c. ad 200–500), and especially during the early and high Middle Ages (c. ad 500–1000), again followed by peaking settlement densities from ad 1000 (Berger, 1996).

Sedimentary history

A simplified overview of the sedimentary history of the Valdaine (Figure 2) is based on published (Berger, 1996, 2003; Berger et al., 2007a) and unpublished data. This overview represented widespread trends, and the local sedimentation and incision history of specific sites may differ from this regional pattern. There are several larger periods of landscape instability with deforested landscapes and/or periods of erosive rainfall, which result in soil erosion and subsequent colluvial and alluvial deposition (Berger, 1996). Landscape stabilization, for instance, related to reforestation or periods with less erosive rainfall events, results in decreased erosion. As a consequence, sediment deposition decreases or even ceases, and in turn, some rivers incise. This is often combined during the Holocene with the formation of soils in alluvial and colluvial deposits (Berger, 2003; Berger et al., 2007a). Also, 20th-century land abandonment resulted in forest increase, soil development, and fluvial incision. Early Holocene floodplain deposits (and records of incision phases) only occur at a few sites, and a detailed history is still partly missing except for the Citelle floodplain (Berger et al., 2002). A regional channel incision period is observed at different sites around 5500 bc. This period is followed by several deposition and stability periods during the Neolithic until the transition with the Bronze Age (c. 5000–2000 bc). An interplay of natural or anthropogenic deforestation and rainfall events determines the severity of local soil erosion and sediment deposition.

Figure 2.

Sedimentary history of the Valdaine region as summarized in available studies. Red lines indicate the different periods of the time-differentiated sediment budget (Figure 5). Symbol size represents rather the duration of a process than the intensity of the process. The temporal resolution used in the original papers amounts to 100–200 years. (a) Floodplain history based on Berger (1996), Thiercelin-Ferber (2012), and unpublished data. (b) Colluvial history based on Berger (1996), including mainly colluvial sites from the central and southern Valdaine. (c) Colluvial history for large dry valleys on the northern hillslopes, based on unpublished data from Brochier (Brochier and Berger, in preparation) and Vital et al. (1999).

During the Bronze Age and the transition phase between Bronze and Iron Age (c. 2000–700 bc), a stability period occurs on the hillslopes and the piedmonts, with local incision of the floodplains. This is followed by a rapid infilling phase during the first centuries of the Iron Age (700–500 bc). The High Roman Empire (ad 1–200/300) is a major soil erosion period, resulting in major colluvial and alluvial sedimentation. Locally, the fluvial system becomes temporarily torrential with the aggradation of bed load deposition by braided rivers. Later on (ad 300–450/500), there is a gradual hillslope stabilization (with local paleosol development), although this is combined with continuing irregular fluvial deposition. A further regional stabilization of the soils occurs between ad 800 and 1200/1300, causing a clear water effect and renewed fluvial incision (Berger, 2003). An important sedimentation period started in the 13–14th century ad, lasting until the beginning of the 20th century, with widespread colluvial deposition and deposition of large amounts of coarse braided river and torrential fluvial deposits. During the 20th century, the upper slopes were reforested and stabilized (Liébault and Piégay, 2001), resulting in clear water effect and fluvial incision of the floodplains. It is however unclear how colluvial deposition reacts, because of the absence of dates for this recent period.

Colluvial sediments vary in nature, depending on the moment of deposition and the source material. They are in general fine grained (silty clay loam to loam), with locally some admixture of gravel. A small fraction is locally more coarse grained (sandy, sometimes with some gravel), especially for deposits from the Roman Period or more recent. Many colluvial deposits from the Bronze Age contain ceramic fragments or other archaeological remains. Overbank alluvial deposits from before ad 1200 are mainly fine grained (silty clay loam), while bed load deposits and deposits from torrential phases during the Roman Time Period are coarse grained (coarse gravel). Alluvial deposits from after ad 1200 exist mainly out of coarse gravels.

Methods

Landscape units

The topography of the Valdaine region is determined by its underlying geological structures. The outer fringes of the Valdaine are pre-mountain limestone ridges, while the lower and less steep parts have a marl underground. Also, the upper parts of the Jabron and Roubion catchments, upstream the Valdaine region, exist of a pre-mountain region with limestone ridges and marl depressions. In the Valdaine, the lower slopes of the pre-mountain ridges are covered by Pleistocene footslope deposits. On the northern slopes of the catchment, a system of Holocene colluvial (dry or perennial) valleys has developed in these Pleistocene deposits. In the southern part of the Valdaine, two limestone ridges occur, with undulating marl colluvial basins (‘cuvettes’) in between.

The central part of the Valdaine is a structural depression with mostly less resistant marl deposits. The topography is less expressed than for the edges and is generally undulating (alternation of small hills and valleys with often a brown to calcareous Pleistocene colluvial soil cover). Some large Pleistocene river terraces occur, which are generally flat or slightly undulating when they are incised after deposition. These terraces are dominated by gravel and are locally covered by loess. Holocene floodplains widen downstream with two major floodplains (Roubion and Jabron rivers). A large alluvial fan is formed at the catchment outlet, just downstream the Jabron–Roubion confluence, and the city of Montélimar is located on this fan.

To compile the sediment budget, we stratified the study area into nine landscape units (Table 2). Average sediment thickness values are calculated for each landscape unit based on available field data, and area-specific values (average deposition per unit area) are then extrapolated for the entire landscape unit. This assumes that these average values are representative for the concerned unit. This methodology has previously been used in other studies (e.g. Macaire et al., 2002; Seidel and Mackel, 2007). The landscape units are based on topography and (Quaternary) geology. They are mapped based on an existing geo-archaeological map (1:25,000; digital version; Verhagen and Berger, 2001), a soil map (Bornand, 1967), topographic data, and additional field verifications. In order to maximize the amount of data for each landscape unit (to have a more reliable average value), landscape units are deliberately kept large. They represent a relatively homogenous regional topography, lithology, and land use history.

Quantification of hillslope deposits

This study deals with the Valdaine region (334 km²) and excludes the upper parts of the Jabron and Roubion catchments (276 km²; thus, total catchment area = 610 km²). These upper parts are a Pre-Alpine mountain region, with landscapes, geomorphology, and topography largely differing from the Valdaine. In addition, quantifiable data on colluvial deposits is lacking for these upper parts.

Hillslope deposition calculations are based on 154 corings and 32 profiles, a large part of which was gathered from existing literature and datasets (for sources, see Table 1). It is difficult to compare data density with other studies, as this study not only relies on coring data but also on profile information. Profiles vary in length from a few meters to 1.115 km. For each hillslope landscape unit, the average sediment depth (SD_unit; m) was calculated as a weighted sediment depth

S D = \frac{\sum_{i = 1}^{n d} S D_{i} W_{i}}{\sum_{i = 1}^{n d} W_{i}}

(1)

Table 1.

Used sources. Each source is shortly described, and the amount of data is mentioned. Different sources provide the same transects.

No.	Reference	Type	Details	Data
1	Berger (1996)	PhD	Profiles through different landscape units; often very long profiles	Ca 65 profiles, 4 corings, 24 deep corings
2	Berger (2011)	Paper	–	10 profiles
3	Berger and Guilaine (2009)	Paper	–	9 profiles
4	Berger and Jung (1999)	Archaeological report	Rescue excavation for TGV rail line	3 profiles (including several corings)
5	Berger et al. (2000)	Paper	Profile; historical and geomorphological information	1 profile
6	Berger, personal communication (2013)	Unpublished data	Original site reports of archaeological and geomorphologic transects, corings and profiles. Personal archive at Université Lyon II Lumière.	(originals of sources 1, 4)
7	Brochier, personal communication (2013)	Unpublished data	Original reports of archaeological rescue excavations and detailed coring descriptions. Personal archive at CAP Valence.	67 corings, 18 profiles (including source 10)
8	Brochier and Berger (in preparation)	Personal archive/publication in preparation	Overview of geomorphologic history of catchment.	(based on sources 1, 4, 6, 7)
9	Ronco et al. (2008)	Archaeological report	Coring data along a long transect	110 corings
10	Thiercelin-Ferber (2012)	Archaeological report	Rescue excavation; profile through floodplain	1 profile
11	Verhagen and Berger (2007)	Paper	Geo-archaeological map of the Valdaine
12	Vital et al. (1999)	Paper	Archaeological excavation; profiles and coring transects through small (colluviated) valleys	1 transect
13	This study	Paper	Additional corings on hillslopes (11) and in floodplains (9). Field mapping.	20 corings

where nd is the number of datapoints for that landscape unit, SD_i is the sediment depth for datapoint i (m), and W_i is the weighting factor for datapoint i (m). Datapoints are profiles, profile pits, or individual corings. The weighting factor W_i is calculated as the length of profile i, with a minimum of 25 m used for the smaller profiles and corings. A value of 25 m is used to take into account spatial autocorrelation, although exact values are unknown for this autocorrelation and might differ between sites. If only a weighting is used for the profiles, and not (a value of 1) for the corings, calculated values would almost completely depend on these profiles. Houben (2008) has demonstrated how erosion and deposition patterns are correlated for inter-coring distances of up to 20 m in the studied loess catchment. Error quantification was performed using Gaussian error propagation (Notebaert et al., 2009). In order to compare the volumes quantified in this study with masses yielded by other studies, we applied a bulk density value of 1.5 g/cm³, comparable with other studies (e.g. Hoffmann et al., 2007).

Floodplains and alluvial fan

The quantification of floodplain deposition is based on detailed mapping of floodplain extent and extrapolation of floodplain sediment thickness data. Similar to other studies (e.g. Notebaert et al., 2011a; Seidel and Mackel, 2007; Stolz, 2011), floodplains were divided in geomorphic units, and for each unit, an average sediment thickness was calculated (equation 1). This is done in order because of the different geomorphology, and related sediment values, between smaller tributaries and for instance the main valleys. The fluvial deposits in the Valdaine region contain a significant but variable amount of recent (post Medieval) gravel layers, which limit coring and digging possibilities. As a consequence, there are only a few data. For the different tributaries, a total of 12 floodplain transects or profiles are available, while 4 transects are available for the main floodplains of the Roubion and Jabron. The floodplain transects exist typically of a combination of smaller profiles and some individual corings to connect these profiles. In addition, for the Roubion, also thickness data from c. 200 m long cutbank in the most upstream stretch in the Valdaine region was available and combined with the transect data.

For the parts of the Roubion and Jabron floodplains upstream the Valdaine, floodplain thickness data were lacking. Floodplain deposition was nevertheless calculated for these upper catchments by measuring the height of the Holocene deposits in cutbanks at sites where the base of the Holocene deposits could be observed. Because of 20th-century river incision (Liébault and Piégay, 2001), cutbanks have incised at several locations down to the underlying bedrock. We assume that the thickness of the fluvial deposits is constant for the entire width of the floodplain.

At the outlet of the catchment near the city of Montélimar, the Roubion and Jabron form a large alluvial fan (6.2 km²). Only a limited number of corings are available for this fan, mainly in the city center of Montélimar, which is built on the apex of the fan. In addition, coring data are available for the Rhône floodplain next to the edge of the fan. Based on the available point data, the volume of fan sediment is calculated in two different ways: (1) first, based on topography, and (2) second, based on a hypothetic cone shape. For the calculation based on topography, a uniform level of the basal fan surface is assumed. Two different assumptions are used to estimate for this level: 68 and 71 m a.s.l. This basal fan level is subtracted from a contemporary digital elevation model (DTM; 25 m resolution). Volume calculation using a theoretical cone shape assumes a uniformly decreasing sediment thickness from the apex toward the edge of the fan. This uses again two assumptions: one with no deposition at the edge and one with 2 m deposition. This value of 2 m equals the thickness of the Holocene Rhône floodplain deposits. These four scenarios are evaluated to get at a best estimate.

Time-differentiated sediment budget

For various sites, depositional ages are available. Here, we use the published sedimentary history (Berger, 1996), their results together with data from other sources (Brochier and Berger, in preparation), radiocarbon dating, buried ceramics and archaeological remains, and stratigraphic correlations (e.g. Berger, 1996). Radiocarbon dating provides maximal ages with an analytical precision associated with the dead of the dated organic remains. Archaeological remains, by contrast, place sediments within a cultural period provided that there was no reworking. Ceramics, for example, from the Roman Period, can often be dated with a better precision than calibrated radiocarbon dates. Age interpretation based on archaeological remains has to be treated with caution (except when buried in situ) and is best used in combination with physical dating methods.

The construction of a temporary differentiated sediment budget requires that sedimentation patterns are quantified for well delineated time periods. Here, time periods are defined based on important catchment wide geomorphologic changes that are present in the sedimentary record (e.g. Trimble, 1999; Verstraeten et al., 2009b). We determined three depositional periods separated by two periods of hillslope stability and local fluvial incision, which represent the major time periods for the sediment budget. The definition of these periods is based on their widespread impact on the sedimentary history for the region, their identification at as many sites as possible, and the involved uncertainties in the identification. At a smaller timescale, deposition periods for one site can correspond with stability or even incision period for another site (Berger, 2003; Berger and Brochier, 2006; Berger et al., 2007a), while the general pattern of the five used periods is valid for each studied site.

For each site (profile, coring, coring transect), the volumes of the present-day sediment are quantified for the three depositional periods, as well as the eroded volumes for the intermediate incision periods (when possible). Eroded volumes for these incision periods are calculated by applying a logical interpolation of the upper surface level of that unit between the points for which this level is still present. This assumes that (1) such a logical, often horizontal, continuation was present, and (2) that for each period, we can define the upper surface level of the original deposits. We take into account only sites where the different depositional units are recognized and the reconstruction of eroded volumes is possible. For each site and each depositional time period (dp), we calculated the relative amount of sediment deposited (SR_dp) as a function of the contemporary total amount of Holocene sediment present (S_Hol) as

S R_{d p} = \frac{S_{d p}}{S_{Hol}}

(2)

with S_dp the sediment deposition during time period p (m³/m²) and S_Hol the contemporary total Holocene sediment deposition (m³/m²). Likewise, the relative erosion amount (ER_ep) for each incision/stability period (ep) was calculated

E R_{e p} = \frac{E_{e p}}{S_{Hol}}

(3)

with E_ep the erosion amount (m³/m²) during incision period ep.

Values for SR and ER are calculated for colluvial and ravine deposits, alluvia in tributaries, and alluvial deposits in the main valley. We did not make a differentiation between the different landscape units because of the too limited amount of data this would yield for some units. They are combined with the total quantities for these different sinks to result in a time-differentiated sediment budget. In order to compare SR and ER values for the time periods with a different length, values are calculated as rates per 1000 years (%/1000 yr). By taking the total sediment deposition into account (see results), these values can be converted to masses per unit surface area (such as 10⁻⁶ m³/m²/yr).

Budget closure

The available data only provide information on sedimentary sinks and fail to explain the entire sediment pathways and how they change over time. Constructing these pathways is challenging for sediment budgets over time frames of centuries to millennia, as data are often lacking and results depend on some fundamental assumptions (e.g. Notebaert et al., 2011a). Quantifications that are based on such fundamental assumptions are, in the sediment budget, displayed in a different way when compared with quantifications that are based on a more solid dataset. Sediment export out of the catchment for periods D2, S2, and D3 (for definition of the periods D1–D3 and S1–S2, see ‘Results’ section) is roughly estimated at 25 × 10³ m³/yr, based on a monitored regional sediment yield of 50–150 t km²/yr (Delmas et al., 2012), a specific weight of 1.5 t/m³, and a catchment area of 610 km². For the early Holocene (D1–S1), this value is corrected in order to have the same ratio for the export as for the main floodplain deposition between periods D1 and D2. This results in an export of 9.4 × 10³ m³/yr.

We estimated the relative contribution of the mountainous headwater catchments (mainly producing coarse sediment) and hilly central parts of the catchment (producing finer sediment) to the fluvial system based on the abundance of gravel layers (described in the profile descriptions). We tentatively used values of 15%, 25%, and 75% for the upstream contribution to sediments in the main floodplains for D1, D2, and D3, respectively, and values of 10%, 15%, and 50% for the tributaries. These values are based on an estimate of gravel amounts in tributary and main valley floodplain deposits based on coring descriptions. Measured amounts of gravel are missing in the original field reports, which results in a large uncertainty on these estimates.

Results

Quantification of Holocene deposition

Based on the different landscape units (Figure 1a, Table 2) and the distribution of sediment thickness in these units (Figure 3), the sediment budget (Figure 4) was created. Total (colluvial) hillslope deposition equals 167.4 ± 13.5 × 10⁶ m³, and 5.2 ± 1.8 × 10⁶ m³ is deposited in (dry) mountain ravines (Table 3). Colluvial deposits cover large areas of the lower parts of the study area. They vary in texture from fine (silty–clayey) to coarse sediment (sand and gravel), and often contain anthropogenic fragments (e.g. bricks, tegulae, ceramics, and (micro) charcoals). Thickness distributions of the colluvial deposits vary significantly between landscape units (Figure 2), using a Kruskal–Wallis Rank Sum Test and a p value of 0.01.

Table 2.

Main properties of the different landscape units. Average slope gradients and standard deviation are calculated based on a digital elevation model with a 50 m pixel resolution using ARCGIS.

Landscape unit	Average slope gradient	Area km²	Geomorphology
1. Bedrock outcrops	23.9 ± 15.2	119 ± 11.9	Bedrock outcrops on limestone ridges and nearby marl terrain.
2. Pleistocene fans	13.8 ± 14.7	22.4 ± 2.4	Pleistocene fans and slope deposits on the northern ridge.
3. Northern dry valleys	5.0 ± 4.6	11.0 ± 1.1	Holocene dry and perennial valleys on the Pleistocene fans.
4. Central zone	3.5 ± 4.2	72.5 ± 7.3	Rather flat central part of the Valdaine; eroded Quaternary river terraces, sometimes covered by loess; low marl hills.
5. Colluvial basins	10.3 ± 10.5	43.0 ± 4.3	Basins formed at the edge of the southern ridges, filled with Pleistocene en Holocene colluvial material.
6. Table terraces	1.9 ± 2.2	23.0 ± 2.3	Quaternary terraces in the central Valdaine zone, non-eroded.
7. Ravines	–	1.9 ± 0.6	Ravines formed in the limestone and marly ridges.
8.a. Tributary floodplains	–	11.3 ± 2.3	Tributary floodplains in the Valdaine region.
8.b. Jabron and Roubion floodplains	–	22.1 ± 1.7	Main floodplains in the Valdaine region
9. Fan	0.9 ± 1.0	6.2 ± 0.6	Alluvial fan of the Roubion and Jabron rivers.
Upstream floodplains	–	7.2 ± 1.4	Floodplains upstream the Valdaine.

Figure 3.

Distribution (weighted histograms, expressed as density) of the thickness of Holocene colluvial deposits (sediment depth in m) of the different landscape units.

Figure 4.

Holocene sediment budget for the Valdaine. Sediment volumes (10⁶ m³) deposited in or eroded from the different sediment sinks are based on available field data (red and orange arrows, where orange arrows indicate sinks/sources with limited data availability). Sediment volumes connecting these sinks and sediment export are based on a number of assumptions and do not correspond with measured quantities (blue errors; see text).

Table 3.

Quantification of Holocene sediment deposits in the Valdaine region and in the upper catchments of the Roubion and Jabron Rivers. For the different landscape units, average sediment thickness values are combined with areas to calculate the volumes of sediment present. Colluvial deposition is only calculated for the Valdaine. Alluvial deposition is calculated for the upstream parts of the Roubion and Jabron catchment as well as for the Valdaine. The last column (n) gives the total number of transects and individual corings available for each sediment sink.

Landscape unit	Area (km²)	Thickness (m)	Volume (10⁶ m³)	n
Pleistocene fans	22.4 ± 2.4	0.51 ± 0.05	11.4 ± 1.7	51
Northern dry valleys	11.0 ± 1.1	1.66 ± 0.15	18.2 ± 2.5	31
Central zone	72.5 ± 7.3	1.22 ± 0.11	88.3 ± 11.7	49
Colluvial basins	43.0 ± 4.3	1.15 ± 0.08	49.4 ± 6.0	55
Stable terraces	23.0 ± 2.3	0	0	–
Total colluvial	–	–	167.4± 13.5	187
Ravines	1.9 ± 0.6	2.79 ± 0.45	5.2 ± 1.8	3
Tributary floodplains	11.3 ± 2.3	2.4 ± 0.3	26.6 ± 3.0	10
Jabron and Roubion floodplains	22.1 ± 1.7	(4.2 ± 1.5)	95.7 ± 36.7	4
Fan	6.2 ± 0.6	–	35 ± 5	8
Total floodplains	39.6± 2.9	–	157.3± 37.2	22
TOTAL sediment deposition Valdaine	334	–	329.9± 39.6
Upstream floodplains	7.2 ± 1.4	2.7 ± 0.5	19.5 ± 3.5
Floodplain deposits entire Roubion and Jabron catchment	610	–	176.8± 37.3

Floodplain deposition was estimated to be 176.8 ± 37.3 × 10⁶ m³, of which 95.7 ± 36.7 × 10⁶ m³ is deposited in the main floodplains (Jabron and Roubion), 35 ± 5 × 10⁶ m³ in the alluvial fan at the outlet, and 26.6 ± 3.0 × 10⁶ m³ in the tributaries, while 19.5 ± 10.5 × 10⁶ m³ is deposited in the upper floodplains. Uncertainty is large on the calculated values of the main valleys because of the limited number of data, even though this uncertainty corresponds only to calculation errors and does not include uncertainty related to representativeness of the available data. Fluvial deposits have a large textural range, from clay to coarse gravels. Coarse deposits (gravel and sand) are typically deposited during active periods, coinciding with torrential waterflow and a braided river system. Sandy to clayey deposits are related to less active fluvial periods, or less active parts of the floodplains. The floodplain deposits show different cut and fill periods. Fluvial deposits typically contain anthropogenic markers (e.g. ceramics, tegulae, and brick fragments), especially from the Roman Period.

The sediment thickness at the apex of the fan, in the city center of Montélimar, varies between 7.2 and 15.2 m, with an average of 11.8 ± 3.0 m (n = 9). The calculated volume varies between 29 and 50 × 10⁶ m³ sediment (Table 4). Taking into account the most appropriate assumptions, the volume can be estimated to fall within 30 to 40 × 10⁶ m³. A best estimate of 35 ± 5 × 10⁶ m³ is used in further calculations. The fan deposits overlay coarse Pleistocene deposits and consist of silt and clay with some gravel layers.

Table 4.

Holocene sediment volume present in the Roubion and Jabron fan near Montélimar, for the different methods and assumptions used (see text).

Method	Sediment deposition (10⁶ m³)
Topographic – Pre-Holocene height 68 m	50
Topographic – Pre-Holocene height 71 m	30
Distance from apex: no deposition at edge	29
Distance from apex: 2 m deposition at edge	41

Holocene and time-differentiated sediment budgets

Based on the sedimentary history, three major depositional periods (D1–D3) separated by two stability and/or fluvial incision periods (S1–S2) are used for the sediment budget (Figures 5 and 6): D1 – early Holocene to Bronze Age (c. 8500–2000 bc); S1 – Bronze Age (c. 2000–700 bc); D2 – Iron Age to early ‘Medieval Period’ (700 bc–ad 800); S2 – the Middle Ages (ad 800–1200); and D3 – the last 800 years (c. ad 1200–2000).

Figure 5.

Time-differentiated sediment budget of the Valdaine Region. Values based on measured field data are given in red, values based on a number of assumptions (see text) are given in blue. Sediment fluxes, sinks and sources are given as rates (10³ m³/yr) in order to correct for the different lengths of the time periods. The sediment budget demonstrated how the different sediment sinks increase in importance over time.

Figure 6.

Average deposition and incision amounts for (a, e) colluvial, (b, f) alluvial, (c, g) and ravine sites of the time-differentiated sediment budget. For colluvial deposition also a budget was created with a finer temporal resolution (d, h; see ‘Discussion’). Values are expressed in % per 1000 yr (a–d), with percentages being the fraction compared with total contemporary storage (see ‘Methods’). These values are converted to catchment area-specific rates (10⁻⁶ m³/m²/yr; e–h). See text for discussion.

In order to connect the different sediment sinks, fundamental assumptions had to be made (see ‘Methods’ section), and results obtained based on these major assumptions are depicted in a separate color in the sediment budget (Figures 4 and 5). The fractions of deposited (and eroded) sediment for each time period is calculated based on 26 colluvial sites, 3 mountain ravine sites, and 4 floodplain sites. Results show that for both colluvial and alluvial sites, sediment deposition is most important during period D3, followed by period D2, while for the mountain ravines period, D2 was the most important period. Erosion by incision was very high for these ravines during period S2, but this calculated value entirely depends on one site (n = 3). Also for alluvial and colluvial sites, S2 was more important than S1.

Discussion

Methods and errors involved

The used methodology is based on the assumption that the sedimentary data are representative for the defined landscape units. Only for the floodplains of the two major rivers, Roubion and Jabron, data density is very low. Consequently, standard errors appear rather high, even if they do not take into account uncertainty resulting from a possible non-representativeness of the data.

The main shortcoming of the presented sediment budget is the unknown sediment export out of the catchment. Sediment export was estimated based on rough regional estimates (which even do not include bed load transport), but catchment-specific data to confirm these values are missing. Fan sediments contain a limited amount of gravel; however, data are lacking to estimate the gravel or bed load fraction for the export. The relative importance of the different sediment sources remains also unknown. Detailed coring which includes a quantification of the gravel fraction are missing, and the quantification of deposited gravel includes a very large uncertainty. In addition, not all gravel originates from the mountainous zones, while not all fine material originates from the central zone. We propose that future research can concentrate on determining the relative importance of both sources by applying a finger print technique and a mixing model (e.g. D’Haen et al., 2012) to the different deposits. In order to deal with these two major uncertainties in the sediment budget, and to avoid that they are interpreted as accurate calculations, we have indicated depending values in a separate color in the budget.

Spatial variation in sediment deposition

Because the main floodplains of the Roubion and Jabron have their source area upstream the Valdaine, it is difficult to compare the alluvial and colluvial deposits based on the values from Table 3, as the catchment area differs for both types of deposits. In order to get better insight into the quantities of sediment deposited in a typical Valdaine subcatchment, separate calculations were made for some selected subcatchments (Figure 1a). These were selected based on the density of field data and to represent the heterogeneity in the Valdaine landscape units. Results (Table 5) show a large variation in area-specific deposition depending on the landscape structure of the subcatchments. Results further show that for subcatchments of the central Valdaine, colluvial deposition is most important. It is also clear that the major fluvial sedimentary sink comprises the lower Roubion and Jabron floodplains.

Table 5.

Holocene sediment budgets for some selected subcatchments. Total amounts of deposited sediment are given (left column, 10⁶ m³), as well as catchment area-specific values (m³/m²).

Subcatchment	Area (km²)	Colluvial deposits		Mountain ravines		Alluvial deposits		Total deposition
		(10⁶ m³)	(m³/m²)	(10⁶ m³)	(m³/m²)	(10⁶ m³)	(m³/m²)	(10⁶ m³)	(m³/m²)
Vermenon	32.1	19.2	0.60	0.1	0.00	9.1	0.28	28.5	0.89
Manson	14.6	14.0	0.96	0.0	0.00	2.7	0.18	16.6	1.14
Ancelle	77.8	45.1	0.58	1.0	0.01	3.2	0.04	49.2	0.63
Citelles	26.4	4.9	0.19	0.9	0.03	4.3	0.16	10.2	0.39
Lanson	6.3	1.3	0.20	0.3	0.05	1.3	0.20	2.9	0.45
Entire Roubion/Jabron	610	??	??	??	??	176.8	0.29	??	??
Valdaine without Roubion/Jabron floodplains	334	167.4	0.50	5.2	0.02	26.6	0.08	199.2	0.60
Upper Roubion/Jabron	276	??	??	??	??	19.5	0.07	??	??

Comparison with other regions

Compared with other Holocene sediment budgets, area-specific colluvial deposition (0.75 × 10⁶ g/m²) is rather high (Figure 7a). Values reported in the European loess region range typically between 0.1 and 0.4 × 10⁶ g/m². For the Valdaine subcatchments, values vary between 0.3 and 1.4 × 10⁶ g/m². Area-specific alluvial deposition for the R/J catchment (0.43 × 10⁶ g/m²) and subcatchments (0.08–0.43 × 10⁶ g/m²) fall within the variation in reported values (Figure 7b). Higher amounts of total sediment deposition point toward higher total Holocene erosion, although sediment export out of the catchment, and thus total erosion, remains unknown.

Figure 7.

Area-specific Holocene (a) colluvial and (b) alluvial sediment deposition during the Holocene for the Valdaine region and some selected subcatchments (Figure 1d), compared with values from other sediment budget studies in Europe. These studies are grouped per country with data from Belgium and a Belgian–Dutch catchment (De Moor and Verstraeten, 2008; Notebaert et al., 2011a, 2013; Verstraeten et al., 2009b), France (Macaire et al., 2002), Germany (Fuchs et al., 2011; Förster and Wunderlich, 2009; Hoffmann et al., 2007; Houben, 2012; Seidel and Mackel, 2007), Italy (Borrelli et al., 2012), and Turkey (Dusar et al., 2011). Masses of sediment (10⁶ g) are divided by the total catchment area to arrive at the area-specific mass (y axis, 10⁶ g/m²).

Time-differentiated budget and sediment cascade model

This simplification of the sedimentary history into three deposition and two erosion periods ignores some stabilization and incision periods for different reasons. A fluvial incision period around 5500 bc is not included as it can only be recognized for a few sites. Another fluvial incision period (c. 200 bc–1 ad) separates the Iron Age deposits from the Roman Period deposits. This period can be separately recognized for 14 sites (out of 32 with a good chronology), 10 of which are stable during this period and 4 with fluvial incision. Because of the lower amount of usable data, it was decided to not make a separation for these periods in the sediment budget. Based on the available data, colluvial deposition is calculated to be 5.2% of total contemporary present sediments (=10.4%/1000 yr or 52 × 10⁻⁶ m³/m²/yr; n = 9) for the Iron Age (700–200 bc) and 19.8% (=24.8%/1000 yr or 124 × 10⁻⁶ m³/m²/yr; n = 7) for the Roman Period and early Middle Ages (ad 1–700/800). These numbers indicate that Roman Period deposition was almost 2.5 times more intense than Iron Age deposition and 10 times higher than early- and mid-Holocene deposition (Figure 6).

An important fluvial incision period which is not separately calculated in the sediment budget is the 20th-century fluvial incision, coinciding with reforestation of the upper catchment (Liébault and Piégay, 2001) and a warming period. This fluvial incision period coincides with ongoing erosion and colluvial deposition in the agricultural areas (Berger, 1996). This period was not separately included in the budget because of the absence of quantitative information in the available dataset. Quantification of incised/eroded amounts in the floodplains would require reconstruction of the pre-incision surface and thus different techniques. For one site on the Roubion, Liébault et al. (2012) have measured a vertical mean incision of 0.5 m for this period. Qualitative field observations for different sites show comparable or even higher height differences between the current lower floodplain and the 19th-century floodplain level. Based on the soil map and topographic data, we calculated the incised area (=current lower floodplain) for the main floodplains to be 1.7 km². A tentative calculation based on these numbers yields an erosion of 0.9 × 10⁶ m³. This is 13%/1000 yr if incision is supposed to have occurred over 100 years, although we have no data for the start of the incision. This incision equals 1.3% of the present-day amount of sediment in the Roubion main floodplain.

The major breaks in the sediment budget coincide also with major cultural breaks. Where these cultural breaks are not necessary sudden events, this is also true for sedimentary changes at a catchment scale, which introduces an error in the sediment budget. Many sedimentary units that can be identified in the field (e.g. Berger, 1996, 2011) can indeed be directly attributed to a specific cultural period. The relation between breaks in the budget and in cultural periods is not only the result from coinciding processes but also from the dating method (using archaeological remains, whose precision is variable in time) and dating resolution. Many sites are dated by a combination of radiocarbon dates and the presence of abundant archaeological material from a particular time period. As a result, sedimentary units are often linked to cultural periods with distinct archaeological traces. The temporal resolution depends on the length of distinguishable cultural periods to which separate sedimentary units can be coupled. Based on the length of the cultural periods and the combined use with radiocarbon dating, the uncertainty on the boundaries of the sediment budget is 100–200 years. Periods of up to 200 years that lack events (no soil erosion or river incision) may not be detected. As a result of these dating constraints, it was, for instance, not possible to distinguish a separate phase of braided river formation during the Roman Period, nor was it possible to determine whether this braided phase occurred synchronously at a regional scale.

Based on the sediment budget, sedimentary history, and sediment properties, a sediment cascade model can be constructed (Figure 8). For this sediment cascade, we hypothesize that the mountainous area produces mainly course sediment which is routed through mountain ravines toward tributaries or directly toward the main Roubion and Jabron floodplains. Sediment eroded in the central part of the catchment is partially deposited on the hillslopes. Another fraction of this sediment is routed through dry valleys to the tributaries and/or directly toward the main Roubion and Jabron floodplains. The field data indicate that reworking of previously deposited sediment is important (see Figure 6), and different colluvial and alluvial sinks have acted for at least part of the Holocene as sediment sources.

Figure 8.

Simplified sediment cascade model for the Valdaine: (a) deposition periods (D1–D2), (b) braided deposition periods (D3), and (c) stability and incision phase (S1–S2). Sediment pathways are schematically represented, with the width of the arrows indicating the importance of the process (they are not quantitatively scaled). The blue/red line represents a topographical cross section from the mountain ridges (blue) toward the central parts of the Valdaine with rolling hills (red). Changes in sediment pathways, sources, and sinks occur between the different depositional periods and the stability/incision periods. During periods D1 and (large parts of) D2, mainly fine sediment is deposited in the floodplains. During period D3 (and some periods of D2), a large amount of coarse sediment is being deposited, originating from the mountainous area and the river type changes to a braided river system. Sediment production in the mountain ridges and the central hills decreases during the stability/incision periods (S1–S2), resulting in incision of the floodplain.

The sediment cascade model indicates how the mountainous region and the central region underwent a change as regards their relative role as sediment sources within the sediment cascade. During depositional phases D1 and D2, mainly fine material is deposited on floodplains (Figure 8), originating mostly from the central part of the Valdaine. Single channel rivers are established, probably with a meandering pattern. During torrential river flow phases, which occur during period D3 (ad 1200–1900) and some parts of period D2 (700 bc–ad 800), large amounts of gravels are deposited by braided rivers in the floodplains (Figure 8), probably mainly originating from the mountainous region. Meanwhile, erosion of fine sediments in the central parts remains high, as is testified by the importance of (fine) colluvial deposition in this zone. During the stability and fluvial incision phases, soil erosion on the hillslopes and in the mountainous area are limited, and a clear water effect is responsible for river incision, locally propagating upstream toward the dry valleys (e.g. Berger, 2003). Most probably, a major part of the sediment produced during these incision phases is exported toward the Rhône.

Relation to climate and land use

When comparing the sediment budget with the settlement history of the Valdaine (Berger, 1996), it is apparent major depositional phases occur during the major land occupation phases (Figure 9). By contrast, stability and fluvial incision occur after a decrease in settlement density. The increase in deposition rates over the Holocene (D1–D2–D3) can only be explained by an important influence of the increasing land use in the catchment and an increase of connectivity between slopes and fluvial systems (e.g. Houben et al., 2013). Not only the temporal evolution gives an indication of the importance of land use as a driving factor but also the spatial pattern. Colluvial sediment deposition is especially important in the central Valdaine, which is the part of the Valdaine with highest historical settlement density (Berger, 1996). Only for phase D3, also, the mountainous area is a major sediment source, but historical settlement data do not yet allow to quantify deforestation in this region on a detailed (temporal and spatial) scale. It is difficult to establish links between the sediment budget and the climatic variability during the Holocene. This is because of the lack of local climatic data, on the one hand, and a major difference in temporal resolution of the sediment budget and the regional climatic data, on the other hand.

Figure 9.

Relation of colluvial and alluvial deposition (this study) with climate (based on lake levels from middle France; Magny, 2004), and settlement density (taken from Berger, 1996). The gray-colored background represents the different deposition (light gray) and stability/incision (dark gray) periods. Note the apparent relationship between intense land use and sediment deposition, especially for 700 bc–ad 2000. A relationship with climate is less clear. For interpretation, take into account the less detailed resolution of the sediment budget data.

At a more detailed temporal scale, the relationship between depositional phases and land use is less obvious, and short phases of intense land use (e.g. in the late Bronze Age) are reported to come without considerable deposition (Berger, 1996, 2003). A direct link between environmental variables such as natural deforestation, anthropogenic land use, and climate variability is reported to be absent at a temporal resolution of decades to centuries. The temporal resolution of a sediment budget and catchment wide processes is not fine enough to confirm this. The sediment budget does indicate, however, the importance of anthropogenic forcings on soils from the Iron Age onward, culminating during the High Roman Empire. This confirms recent results obtained from lake archives in the Upper Rhône catchment (Doyen et al., 2013; Simmonneau et al., 2013) and quantitative estimations of soil erosion in small catchment of the Rhine area (Houben et al., 2013).

When considering the temporal resolution of the sediment budget, the peak in sediment deposition during period D3 coincides both with a peak in land use and with a marked climatic period (the ‘Little Ice Age’ (‘LIA’)). In southern France, the ‘LIA’ resulted in an increase in peak floods and droughts (Grove, 2001), in erosion in the high headwaters in the crystalline massifs (Arnaud et al., 2012), and in human pressure on the landscape which resulted in the establishment of braided rivers in the Rhône catchment (e.g. Astrade et al., 2011; Bravard, 2010). Likewise, previous research has shown how braided rivers developed in the French (Pre-)Alps because of deforestation, increased sediment production because of climatic change, or a combination of both (e.g. Bravard, 2010; Liébault and Piégay, 2001). The 20th-century river incision following this braided phase is a result of less (intense) rainfall, of warming conditions at higher altitude, and of land abandonment and reforestation (e.g. Liébault and Piégay, 2001). Our sediment budget confirms the importance of the mountainous areas as a sediment source and the establishment of a braided river during the ‘LIA’ (period D3). The data, however, do not indicate whether this unprecedented erosion period in the mountains is the result of passing a precipitation threshold (as hypothesized by Arnaud et al., 2012), lasting deforestation of the mountainous region, or both. A refinement of historical land use data, especially for the mountain region, combined with more detailed sedimentation rates and climatic data, could help to decipher the relative importance of the environmental drivers.

Conclusion

In this paper, we have presented a sediment budget for the Valdaine Region, located at the edge of the French Pre-Alps. The budget quantifies volumes of Holocene colluvial and alluvial deposition and is differentiated in time with three deposition periods, separated by two stability/incision periods. Overall, the amount of Holocene colluvial deposits is rather high compared with other catchments, while the alluvial deposition falls within the variation found in literature for other catchments.

High rates of sediment deposition (and related soil erosion) correlate well with intense occupation phases from the Late Holocene period (especially the Roman Period and Modern Times); phases of hillslope stability and river incision correlate with periods of decreased human occupation. Human occupation of the region was rather intense since (late) Neolithic times, which may at least partially explain the high amounts of Holocene colluvial deposits found in the Valdaine. It is difficult to demonstrate a climatic influence on the temporal variation in sediment deposition, but the major deposition period (D3, ad 1200–2000) coincides both with a more erosive climate (‘LIA’) as with a major occupation period.

The strengths of the sediment budget is the combination of quantification techniques with a spatial and temporal variation in sediment sinks, which allows to go beyond single site sedimentary deposition rate studies toward a catchment integrated approach. Although our sediment budget has a smaller temporal resolution than for instance lake studies, results are quantified, spatially integrated, and take into account the different sediment pathways. Major future challenges to improve this budget are related to the spatial and temporal resolution. The temporal resolution depends on the possibility to recognize well constraint sedimentation (or incision) phases for a large number of sites in order to calculate catchment wide trends. The combination of radiocarbon dating with dating based on archaeological remains in this study has allowed already to recognize five to seven different stages in the sediment budget.

In this study, a major challenge has been the quantification of sediment export out of the catchment for defined time periods. These values have an important influence on the reconstructed catchment’s sediment flows. In addition, future work should try to integrate fingerprinting in order to get less uncertainty in the quantification of sources.

The sediment budget shows the possibilities and constraints of field based studies to understand variations in soil erosion and sediment flows over a Holocene timescale within a catchment. Modeling of soil erosion and sediment transport might provide a solution for the most important constraints of the approach presented in this study (e.g. Notebaert et al., 2011b). For this study area, it would, however, depend on models which are able to both incorporate mass processes, hillslope process, and fluvial processes, taking into consideration erosion, deposition, and incision, and this for mountainous and undulating landscapes. This also requires detailed and quantitative information on the Holocene of all driving factors, which is not available for the Valdaine.

Footnotes

Acknowledgements

The authors wish to thank H. Piégay (ENS de Lyon, France) for his valuable advice and suggestions, which were the onset of their research. We are also grateful to everybody who helped us finding unpublished and published data, especially the CAPRA-Valence, Frédérique Thiercelin-Ferber (INRAP), and all the archaeologists of the TGV-Méditerranée fast railway fieldwork operation, supervised by INRAP. The authors like to thank P. Houben and one anonymous reviewer for their valuable comments.

Funding

This work was supported by the Research Foundation Flanders – FWO (personal grant and travel grant for Bastiaan Notebaert) and the ‘APR 2012 du Réseau des Observatoires Hommes-Milieux vallée du Rhône’ project, Institut Ecologie et Environnement – INEE (CNRS, France).

References

Arnaud

Révillon

Debret

. (2012) Lake Bourget regional erosion patterns reconstruction reveals Holocene NW European Alps soil evolution and paleohydrology. Quaternary Science Reviews 51: 81–92.

Astrade

Jacob-Rousseau

Bravard

J-P

. (2011) Detailed chronology of mid-altitude fluvial system response to changing climate and societies at the end of the ‘Little Ice Age’ (Southwestern Alps and Cévennes, France). Geomorphology 133: 100–116.

Berger

J-F

(1996) Le cadre paléogéographique des occupations du bassin valdainais (Drôme) à l’Holocène. Paris: Université Paris I, p. 325.

Berger

J-F

(2003) Les étapes de la morphogenèse holocène dans le sud de la France. In: van der Leeuw

Favory F and Fiches J-L (eds) Archéologie et systèmes socio-environnementaux: Études multiscalaires sur la vallée du Rhône dans le programme Archaeomedes. Paris: CNRS, pp. 87–167.

Berger

J-F

(2011) Hydrological and post-depositional impacts on the distribution of Holocene archaeological sites: The case of the Holocene middle Rhône River basin, France. Geomorphology 129: 167–182.

Berger

J-F

Brochier

J-L

(2006) Les apports de la géoarchéologie à la connaissance des paysages et des climats de l’époque médiévale en moyenne vallée du Rhône. In: Maufras

(ed.) Habitats, nécropoles et paysages dans la moyenne et basse vallée du Rhône (VIIe–XVe s.): Conributions des travaux archéologiques sur le tracé du TGV Méditerranée à l’étude des sociétés rurales médiévales, Documents d’Archéologie française. Paris: Éditions de la maison des Sciences de l’Homme, pp. 163–208.

Berger

J-F

Guilaine

(2009) The 8200 cal BP abrupt environmental change and the Neolithic transition: A Mediterranean perspective. Quaternary International 200: 31–49.

Berger

J-F

Jung

(1999) Rapport fossés: Fossés et voirie. Operation transversalle. AFAN-TGV Ligne 5 Secteur II: Valance-Avignon. Archéologie et TGV: rapport de fouille. Bron: AFAN TGV.

Berger

J-F

Brochier

J-L

Vital

. (2007a) Nouveau regard sur la dynamique des paysages et l’occupation humaine à l’âge du Bronze en moyenne vallée du Rhône. In: Richard

Mordant

(eds) Le Temps, Actes du 109e congrès des CTHS. Emprises, déprises et rythmes agricoles à l’Âge du Bronze. Besançon: Éditions du CTHS, pp. 260–283.

10.

Berger

J-F

Delhon

Bonté

. (2002) Paléodynamique fluviale, climat, action humaine et évolution des paysages du bassin versant de la Citelle (moyenne vallée du Rhône, Drôme) au cours de l’Atlantique ancien (8000-6000 BP) à partir de l’étude de la séquence alluviale d’Espeluche-lalo. In: Bravard

Magny

(eds) Les fleuves ont une histoire. Paléoenvironnement des rivières et lacs français depuis 15000 ans. Paris: Ed. Errance, pp. 223–238.

11.

Berger

J-F

Magnin

Thiébault

. (2000) Emprise et déprise culturelle à l’Age du Bronze: L’exemple du Bassin Valdainais (Drôme) et de la moyenne vallée du Rhône. Bulletin de la Société préhistorique française 97(1): 95–119.

12.

Berger

J-F

Nuninger

Van der Leeuw

(2007b) Modeling the role of resilience in socio-environmental co-evolution: The middle Rhone valley between 1000 bc and ad 1000. In: Kolher

Van der Leeuw

(eds) Modeling Socioecological Systems. Santa-Fé, NM: Sarre editions, pp. 41–59.

13.

Blanchet

(1990) Régimes météorologiques et diversité climatique dans l’espace rhonalpin [Meteorological regimes and climatic diversity in the Rhône-Alpes area]. Revue de géographie de Lyon 65: 106–117.

14.

Bornand

(1967) Etude pédologique du Bassin Valdainais, 1: 25,000. Montpellier: Institut National de Recherche Agronomique, Service d’Etude des Sols, Centre de Recherche Agronomique du Midi.

15.

Borrelli

Hoelzmann

Knitter

. (2012) Late quaternary soil erosion and landscape development in the Apennine region (Central Italy). Quaternary International 312: 96–108.

16.

Bravard

J-P

(2010) Discontinuities in braided patterns: The River Rhone from Geneva to the Camargue delta before river training. Geomorphology 117: 219–233.

17.

Brochier

Berger

(in preparation) Des paysages et du climat de la fin des temps glaciaires à nos jours en moyenne vallée du Rhône d’après les données des travaux archéologiques du TGV Méditerranée.

18.

Delmas

Cerdan

Cheviron

. (2012) Sediment export from French rivers to the sea. Earth Surface Processes and Landforms 37: 754–762.

19.

De Moor

JJW

Verstraeten

(2008) Alluvial and colluvial sediment storage in the Geul River catchment (The Netherlands): Combining field and modelling data to construct a Late Holocene sediment budget. Geomorphology 95: 487–503.

20.

D’Haen

Verstraeten

Degryse

(2012) Fingerprinting historical fluvial sediment fluxes. Progress in Physical Geography 36(2): 154–186.

21.

Dotterweich

(2008) The history of soil erosion and fluvial deposits in small catchments of central Europe: Deciphering the long-term interaction between humans and the environment: A review. Geomorphology 101: 192–208.

22.

Doyen

Vannière

Berger

J-F

. (2013) Land-use changes and environmental dynamics in the upper Rhone valley since Neolithic times inferred from sediments in Lac Moras. The Holocene 23: 961–973.

23.

Dusar

Verstraeten

Notebaert

. (2011) Holocene environmental change and its impact on sediment dynamics in the Eastern Mediterranean. Earth-Science Reviews 108: 137–157.

24.

Förster

Wunderlich

(2009) Holocene sediment budgets for upland catchments: The problem of soilscape model and data availability. Catena 77: 143–149.

25.

Fuchs

Will

Kunert

. (2011) The temporal and spatial quantification of Holocene sediment dynamics in a meso-scale catchment in northern Bavaria, Germany. The Holocene 21: 1093–1104.

26.

Grove

(2001) The ‘Little Ice Age’ and its geomorphological consequences in Mediterranean Europe. In: Ogilvie

AEJ

Jónsson

(eds) The Iceberg in the Mist: Northern Research in Pursuit of a ‘Little Ice Age’. New York: Springer, pp. 121–136.

27.

Hinderer

(2012) From gullies to mountain belts: A review of sediment budgets at various scales. Sedimentary Geology 280: 21–59.

28.

Hirsch

Vincent

(1966) Aspects climatiques du bassin du Jabron (Drôme). Revue de géographie alpine 54(3): 457–467.

29.

Hoffmann

Erkens

Cohen

. (2007) Holocene floodplain sediment storage and hillslope erosion within the Rhine catchment. The Holocene 17: 105–118.

30.

Hoffmann

Glatzel

Dikau

(2009) A carbon storage perspective on alluvial sediment storage in the Rhine catchment. Geomorphology 108: 127–137.

31.

Hoffmann

Schlummer

Notebaert

. (2013) Carbon burial in soil sediments from Holocene agricultural erosion, central Europe. Global Biogeochemical Cycles 27(3): 828–835.

32.

Houben

(2008) Scale linkage and contingency effects of field-scale and hillslope-scale controls of long-term soil erosion: Anthropogeomorphic sediment flux in agricultural loess watersheds of Southern Germany. Geomorphology 101: 172–191.

33.

Houben

(2012) Sediment budget for five millennia of tillage in the Rockenberg catchment (Wetterau loess basin, Germany). Quaternary Science Reviews 52: 12–23.

34.

Houben

Schmidt

Mauz

. (2013) Asynchronous Holocene colluvial and alluvial aggradation: A matter of hydrosedimentary connectivity. The Holocene 23: 544–555.

35.

Liébault

Piégay

(2001) Assessment of channel changes due to long-term bedload supply decrease, Roubion River, France. Geomorphology 36: 167–186.

36.

Liébault

Lallias-Tacon

Cassel

. (2012) Long profile responses of Alpine braided rivers in SE France. River Research and Applications 29(10): 1253–1266.

37.

Macaire

J-J

Bellemlih

Di-Giovanni

. (2002) Sediment yield and storage variations in the Négron River catchment (south western Parisian Basin, France) during the Holocene period. Earth Surface Processes and Landforms 27: 991–1009.

38.

Magny

(2004) Holocene climate variability as reflected by mid-European lake-level fluctuations and its probable impact on prehistoric human settlements. Quaternary International 113: 65–79.

39.

Notebaert

Verstraeten

(2010) Sensitivity of West and Central European river systems to environmental changes during the Holocene: A review. Earth-Science Reviews 103: 163–182.

40.

Notebaert

Verstraeten

Houbrechts

. (2013) Holocene floodplain deposition and scale effects in a typical European upland catchment: A case study from the Amblève catchment, Ardennes (Belgium). The Holocene 23(8): 1184–1197.

41.

Notebaert

Verstraeten

Rommens

. (2009) Establishing a Holocene sediment budget for the river Dijle. Catena 77: 150–163.

42.

Notebaert

Verstraeten

Vandenberghe

. (2011a) Changing hillslope and fluvial Holocene sediment dynamics in a Belgian loess catchment. Journal of Quaternary Science 26: 44–58.

43.

Notebaert

Verstraeten

Ward

. (2011b) Modeling the sensitivity of sediment and water runoff dynamics to Holocene climate and land use changes at the catchment scale. Geomorphology 126: 18–31.

44.

Reid

Dunne

(2005) Sediment budgets as an organizing framework in fluvial geomorphology. In: Kondolf

Piégay

(eds) Tools in Fluvial Geomorphology. Chichester: John-Wiley & Sons Ltd, pp. 463–500.

45.

Ronco

Brochier

J-L

Ferber

. (2008) Rapport de diagnostic avril 2008. Canalisation de transport de gaz. Bron: INRAP.

46.

Seidel

Mackel

(2007) Holocene sediment budgets in two river catchments in the Southern Upper Rhine Valley, Germany. Geomorphology 92: 198–207.

47.

Simonneau

Doyen

Chapron

. (2013) Holocene land-use evolution and associated soil erosion in the French Prealps inferred from Lake Paladru sediments and archaeological evidences. Journal of Archaeological Science 40: 1636–1645.

48.

Slaymaker

(2003) The sediment budget as conceptual framework and management tool. Hydrobiologia 494: 71–82.

49.

Stolz

(2011) Budgeting soil erosion from floodplain sediments of the central Rhenish Slate Mountains (Westerwald), Germany. The Holocene 21: 499–510.

50.

Thiercelin-Ferber

(2012) Route d’Espeluche Aménagement et extension du stade Beaulieu. Auvergne: Inrap Rhône-Alpes.

51.

Trimble

(1999) Decreased rates of alluvial sediment storage in the coon creek basin, Wisconsin, 1975–93. Science 285: 1244–1246.

52.

Trimble

(2009) Fluvial processes, morphology and sediment budgets in the Coon Creek Basin, WI, USA, 1975–1993. Geomorphology 108: 8–23.

53.

Van Oost

Verstraeten

Doetterl

. (2012) Legacy of human-induced C erosion and burial on soil-atmosphere C exchange. Proceedings of the National Academy of Science of the United States of America 109: 19492–19497.

54.

Verhagen

Berger

J-F

(2001) Predictive modelling of buried archaeological sites in the Tricastin-Valdaine Region (Middle Rhone Valley, France). In: CAA 2000: ‘Computing Archaeology for Understanding the Past’, Joint CAA/UISPP Commission IV Conference (eds Stancic

Veljanovic

), Ljubjana, 18–20 April 2000, pp. 219–232. Ljubjana: BAR International Series.

55.

Verhagen

Berger

J-F

(2007) The hidden reserve. Predictive modelling of buried archaeological sites in the Tricastin-Valdaine region (Middle Rhône Valley, France). In: Verhagen

(ed.) Case Studies in Archaeological Predictive Modeling. The Netherlands: Leiden University Press, pp. 41–66.

56.

Verstraeten

Lang

Houben

(2009a) Human impact on sediment dynamics: Quantification and timing: Introduction. Catena 77: 77–80.

57.

Verstraeten

Rommens

Peeters

. (2009b) A temporarily changing Holocene sediment budget for a loess-covered catchment (central Belgium). Geomorphology 108: 24–34.

58.

Vital

Brochier

J-L

Durand

. (1999) Roynac le Serre 1 (Drôme): Une nouvelle séquence holocène en Valdaine et ses occupations des Âges des métaux. Bulletin de la Société préhistorique française 96(2): 225–240.