Search

Trabajo presentado en TERRAenVISION Environmental Issues Today: Scientific Solutions for Societal Issues, celebrado en Barcelona (España) del 27 de enero al 2 de febrero de 2018, This contribution results from international collaborative efforts of different research groups around the Mediterranean basin focusing on soil conservation and water management. We will present the most complete database of runoff and soil erosion information and analyze the records from 68 locations, including more than 22458 detail events between 1985-2015. Soil conservation and water planning are two of the most challenging problems around the Mediterranean basin due to climate conditions and human activity. We believe that future advances on understanding soil degradation by water should be developed under global approaches. In this contribution, we will present an analysis of atmospheric conditions, expressed by weather types and the resulted rainfall, runoff, and erosion and sediment yield around the Mediterranean basin. The weather types compile daily information about the different air masses responsible for rainfall, runoff and eventually erosion and sediment yield, and their analyses let us know the synchrony of the response at different sites around the Mediterranean basin. The analyses of the global dataset display different spatial patterns for rainfall, runoff, and sediment yield, related to different atmospheric patterns. This research will be a valuable tool for understanding the evolution of these environmental variables, and therefore it will allow future planning to design regional water management and soil conservation measures

International audience

International audience; Les observatoires de long terme en environnement permettent d'acquérir une grande quantité de données. Se posent alors les questions de leur bancarisation, leur gestion par les équipes productrices, et leur mise à disposition sur Internet. Le projet BDOH (base de données pour les observatoires en hydrologie) vise à développer pour cela un outil performant, pour les données issues des observatoires hydrologiques d'Irstea. BDOH permet notamment: la visualisation graphique et le téléchargement en temps réel des chroniques, l'import et export de données sous forme de fichiers texte à différents formats avec possibilité d'interpolation du pas de temps, le calcul automatique de données indirectes via des lois de transformation de type courbes de tarage, la gestion de plusieurs niveaux d'utilisateurs avec des accès et droits différenciés, une interface cartographique conviviale.Le projet implique 4 équipes / unités de recherche d'Irstea associées au pôle Informatique Scientifique de la Direction des systèmes d'information, qui assure le développement informatique. Une pré-version est actuellement en cours de test.

Soil erosion is one of the environmental problems of main concern in Mediterranean areas because of significant both on- and off-site issues such as soil loss, land degradation, reservoir siltation, water quality and ecological impacts. Even if it has longly been observed that major erosive events are almost always associated to extremely large or intense rainfall events, little is known on the exact nature of these events at the small catchment scale. One of the main reasons for this is that, frequently, the investigation of variables influencing sediment yield is done through the analysis of erosive events of highly variable magnitude, with the objective of deriving general log-relationships between sediment yield and hydrological variables. With such approach, the specificity of the most erosive events is often lost in the analysis where large events act as a single cluster at one extreme of the events distribution and their real magnitude is masked by the log scale.With the objective of determining the specific characteristics of the most erosive events, and to infer some of the key processes acting during them, this study presents an analysis of the 10 most erosive events recorded for a set of 10 small research catchments located in the Mediterranean region. The different catchments, grouped in the R-OSMed Network, are located in France (4), Spain (3), Italy (1), Portugal (1) and Tunisia (1); the catchments have areas ranging from 0.018 to 1.32 km2, mean annual precipitation from 236 to 1303 mmyear-1 and mean annual sediment yield between 7.5 and 6900 Mgkm-2year-1. In total more than 120 years of hydrological and sediment data (series between 3 and 29 years long) have been analysed to select the 10 most erosive events for each catchment. The study is based on the analysis of the relationship existing between a series of meteorological, hydrological and sediment related variables. First results show that the cumulated sediment yield of these 10 most erosive events represented a variable proportion (from 150 to 1500%) of the mean annual suspended sediment yield. The size of the catchments, their relative area with intense erosion and the mean annual rainfall were identified as the main causes of the differences between the catchments responses. For major erosive events, runoff depth was more dependent on rainfall depth than on rainfall intensity, whereas peakflow was not clearly related to any specific variable. Suspended sediment load during major erosive events was most often related to runoff depth and to peakflow discharge, but considering all catchments, no general relationship could be observed between suspended sediment concentration (maximum and mean) and rainfall or runoff variables. Results also show that the return periods associated to rainfall and flood characteristics may be used as a measure of the singularity of the diverse variables, allowing a better characterisation of major erosive events. Finally, the comparison between the responses of the different catchments helps to improve the knowledge of the hydrological and geomorphological functioning of each of them.

D’initiative française, le projet international HyMeX a pour objectif d’améliorer la compréhension du cycle de l’eau en Méditerranée, de sa variabilité, de l’échelle de l’événement météorologique aux échelles saisonnières et interannuelles, et de ses caractéristiques sur une décennie, dans un contexte de changement global. Le projet est motivé par le rôle déterminant des processus de mésoéchelle, couplés entre l’atmosphère, la mer et la terre, sur la variabilité du système climatique et sur le déclenchement d’événements hydrométéorologiques extrêmes (précipitations et inondations, vents forts et convection océanique, canicules et sécheresses). Le projet vise enfin à évaluer les conséquences de ces événements extrêmes sur la vulnérabilité sociale et économique de cette région et sa capacité d’adaptation.

International audience; On the highly erodible catchments of Draix (French Alps), concentrations of suspended sediments up to 900 g/L have been observed during floods. To investigate the behavior of such flows, an experimental study is conducted in a flume with suspensions of natural sediments at different concentrations. Surface observations suggest that with increasing concentration, the flow evolves from a turbulent to a laminar regime, with surprisingly stable surface patterns at high concentrations. Velocity profiles for the lower concentrations are logarithmic, whereas at higher concentrations, they show a clear unsheared zone that is haracteristic of a yield stress fluid behavior. The thickness of this unsheared zone is used to estimate the yield stress of the hyperconcentrated suspensions. According to these experiments, there is a transition in the flow behavior for concentrations higher than a critical value which is around 600 g/L. This suggests that discharge measurements should be reevaluated for floods of high concentration in Draix streams.

International audience; Conférence invitée lors du International Workshop on Monitoring Bedload and Debris-Flows in Mountain Basins, Free University of Bolzano, Italy, 10-12 October 2012.

International audience

International audience; A key issue in the study of the carbon cycle is constraining the stocks and fluxes in and between C-reservoirs. Among these, the role and importance of fossil organic carbon (FOC) release by weathering of outcropping sedimentary rocks on continental surfaces is still debated and remains poorly constrained. Our work focuses on FOC fluxes due to chemical and mechanical weathering of marls in two experimental watersheds with typical badlands geomorphology (Draix watersheds, Laval and Moulin, Alpes de Haute Provence, France). Organic matter from bedrock, soil litter and riverine particles are characterized by Rock-Eval 6 pyrolysis. FOC fluxes due to mechanical weathering are then estimated by monitoring the annual particulate solid exports at the outlets of the watersheds (1985-2005 period). FOC fluxes from chemical weathering were calculated using Ca2+ concentrations in dissolved loads (year 2002) to assess the amount of FOC released by the dissolution of the carbonate matrix. Results show that FOC delivery is mainly driven by mechanical weathering, with a yield ranging from 30 to 59 t km-2 yr-1 in the Moulin (0.08 km2) and Laval (0.86 km2) catchments, respectively, (1985-2005 average). The release of FOC attributed to chemical weathering was 2.2 to 4.2 t km-2 for the year 2002. These high FOC fluxes from badlands are similar to those observed in tectonically active mountain catchments. At a regional scale, badland outcropping within the Durance watershed does not exceed 0.25% in area of the Rhone catchment, but could annually deliver 12 000 t yr-1 of FOC. This flux could correspond to 27% of the total particulate organic carbon (POC) load exported by the Rhone River to the Mediterranean Sea. At a global scale, our findings suggest that erosion of badlands may contribute significantly to the transfer of FOC from continental surfaces to depositional environments.

National audience; Les rivières de montagne sont réputées pour produire des crues rapides et violentes aux conséquences parfois dramatiques. Afin d'éviter des drames, tels celui du Grand Bornand en 1987, ou celui de la vallée de la Maurienne en 2008, il est impératif d'améliorer la prédiction des crues et de leurs conséquences hydro-géomorphologiques (inondations, érosion des berges, engravements, défluviations). Les crues torrentielles et le transport des sédiments associés contribuent à la dynamique géomorphologique des plaines alluviales et jouent à ce titre un rôle moteur sur l'écologie des hydro-systèmes alpins. Les conséquences de l'interruption ou de l'amplification du transport solide peuvent être à l'origine de dégradations environnementales importantes : disparition des habitats, colmatage des fonds, dépérissement des forêts alluviales.

International audience; In the Southern French Alps, the Black Marls formation, or "Terres Noires" in French, covers a large area. This formation is very susceptible to weathering and erosion. Under the Mediterranean and mountainous climate with frost in winter and high intensity rainfall in summer, it results in "badlands" topography with steep and dissected slopes. The solid transport during floods is very high, both for suspension and bedload transport. The experimental site of Draix was created in the period 1983-1984 and aim principally at studying erosion processes and flood generation in this area. Six watersheds, from 1 000 m2 to 20 km2 are equipped to study and quantify runoff and erosion processes, according to the vegetation cover and basin size. Seven rain gauges record the precipitation in different locations of the area, ranging from 850 to 2200 m in elevation. Discharge is measured in calibrated sections for the 5 smaller watersheds (up to 1 km2) and since 2008 in natural section for the largest (The Bouinenc river). The suspended sediment yield is sampled in the gauging sections and monitored continuously with an optical fiber sensor. Except for the Bouinenc, the bedload sediment yield is measured in a sediment trap, just upstream the gauging stations. The analyses of the 25 year datasets for runoff and erosion in the 5 small badlands catchments pointed out the non-linearity of the response to a rainfall input and the scale effects in sediment production. Temporal and spatial patterns are presented and discussed. The recent monitoring of a larger catchment (The Bouinenc) allows the analysis, for a few events, of the sediment fluxes transferred downstream of such a badlands area.

International audience; In the Southern French Alps, the Black Marls formation, or "Terres Noires" in French, covers a large area. This formation is very susceptible to weathering and erosion. Under the Mediterranean and mountainous climate with frost in winter and high intensity rainfall in summer, it results in "badlands" topography with steep and dissected slopes. The solid transport during floods is very high, both for suspension and bedload transport. The experimental site of Draix was created in the period 1983-1984 and aim principally at studying erosion processes and flood generation in this area. Six watersheds, from 1 000 m2 to 20 km2 are equipped to study and quantify runoff and erosion processes, according to the vegetation cover and basin size. Seven rain gauges record the precipitation in different locations of the area, ranging from 850 to 2200 m in elevation. Discharge is measured in calibrated sections for the 5 smaller watersheds (up to 1 km2) and since 2008 in natural section for the largest (The Bouinenc river). The suspended sediment yield is sampled in the gauging sections and monitored continuously with an optical fiber sensor. Except for the Bouinenc, the bedload sediment yield is measured in a sediment trap, just upstream the gauging stations. The analyses of the 25 year datasets for runoff and erosion in the 5 small badlands catchments pointed out the non-linearity of the response to a rainfall input and the scale effects in sediment production. Temporal and spatial patterns are presented and discussed. The recent monitoring of a larger catchment (The Bouinenc) allows the analysis, for a few events, of the sediment fluxes transferred downstream of such a badlands area.

National audience; L'étude traite de l'évaluation de la qualité altimétrique de modèles numériques de terrain (MNT) construits à partir de mesures LiDAR sur un terrain très accidenté sans végétation. Elle cherche à quantifier l'erreur affectant la mesure LiDAR et la précision attendue sur un MNT dérivé. Cette étude fait également le bilan de la densité de mesures LiDAR utile à la description du terrain et compare la qualité altimétrique de trois MNT. Les conclusions amèneront à la validation de l'un d'eux pour des études sur les processus hydrologiques et érosifs Ces trois MNT sont issus de filtrages de deux nuages de points, qui diffèrent par une recalibration du système avant leur création. Toutefois la méthode de filtrage en triangulation itérative est la même mais utilisant des paramètres différents. La qualification montre que la méthode de filtrage, influencée par les points bas, sous-estime en moyenne les altitudes. La densité de points, renforcée pour un des MNT permet de négliger ce biais et de lui conférer une qualité tout à fait satisfaisante. En effet ce dernier MNT présente un biais de 1.8 cm et un écart-type de 15.1 cm en moyenne. La distribution des écarts altimétriques est ensuite analysée selon la forme du terrain et selon sa pente. Ainsi l'erreur altimétrique augmente sensiblement pour des pentes supérieures à 30°, la précision des MNT est inférieure à 10 cm sur les plats mais peut dépasser 20 cm sur des crêtes ou versants. La qualité obtenue sur le MNT retenu correspond aux attentes des chercheurs utilisateurs. Ce MNT ne permettra cependant pas de quantifier et suivre l'érosion d'année en année. Enfin, une piste d'amélioration possible serait d'adapter les techniques de filtrage à la présence de végétation et à ce type de relief.

International audience; In the highly erosive marly catchments of Draix (Southern Alps, France), downstream fining of sediments has been observed and can not be explained by selective sorting. Moreover, high concentrations of suspended fine sediment (up to 800 g/L) are measured during flood events in these basins. These observations lead to the hypothesis that abrasion and fragmentation of marly sediments during transport play an important role in the production of fine sediments. In order to quantify these processes, experiments are realized in two different devices and provide abrasion rates ranging from 2 to 20%/km. Field measurements show a clear decrease in marl proportion and grain size. A simple model allows us to extract field integrated abrasion rates which appear to be higher than the experimental ones. This might be the evidence of the role of long term in situ degradation processes.

International audience; On marly eroded terrains of the French Southern Alps, many researches are undertaken in order to better understand the role of vegetation and bioengineering works on erosion and sedimentation control. These researches in particular made it possible to develop tools of ecological engineering bound for the practitioners in order to conceive operations for mitigation of damage related to soil erosion. In particular they are methods of diagnosis and strategies for action with bioengineering techniques. These tools must make it possible to guide the choice of the gullies to be rehabilitated and that of the types of works of vegetalisation to be used, in particular via the establishment and the use of a gully typology. Before passing to phases of real use of these tools through expertise, as this is today considered on the scale of the large catchment area of the Durance in France (4000 km²), it appeared convenient to carry out a life-size test of application of these tools. This test was carried out on the marly catchment of Francon (73 ha), which belongs to the experimental complex of Draix (04), labellized Observatoire of Research in Environment (ORE) and of which the objectives are to improve knowledge on the formation of floods and bedload transport in small mountainous marly catchments. On this basin, 30 gullies, representing a total surface area of approximately 20 ha, were thus identified like ecologically suitable for rehabilitation, i.e. on which it appeared possible and convenient to install bioengineering works. This test thus made it possible to check the relevance of the tools proposed to apply an action with bioengineering. An ecological operation of rehabilitation of this basin, carried out jointly with the French Office National des Forêts (ONF)', was then carried out in April 2008 in accordance with the test results. It consisted of the construction of 672 bioengineering works, namely of brush layers and brush mats of cuttings on deadwood microdams, implementing in particular the use of 25.000 cuttings of willows (Salix purpurea and S. incana). Beyond its role of validation of former results, this device is used as experimental site in order to improve the effectiveness of the bioengineering works for the sustainable retention of sediment, in particular by checking their resistance (damage on the works) and their performance (resprout and survival of the cuttings, sediment trapping) during extreme climatic disturbances. In particular, it must make it possible to study the effects of changing of spatial scales, still badly perceived, by observing if the effectiveness of local actions with bioengineering (inside small gullies of 1 ha) indeed reflects the reduction of sediment yield at the exit of larger catchments (about the km²). For this, measurements are made on 363 works in 26 gullies. The first results (2008) especially reveal: i/ a very good resistance of the works: 19 % of the works are partially damaged on less than 1/3 of their surface area, the highest rainfall intensity being 28 mm/h during 1 h; ii/ a total rate of cutting resprout of 97% (99% for S. purpurea and 94% for S. incana), the total precipitation of 2008 being 803 mm; iii/ an average trapping of 0,1 m3 of sediment per work, the latter yet still little filled. Lastly, this catchment area will also serve as demonstration site with the attention of the practitioners, managers and decision makers of the actions of erosion and sedimentation control.

National audience; Les MNT à très haute résolution spatiale sont désormais accessibles pour des secteurs fortement disséqués comme les badlands. Ces représentations numériques de l'altimétrie sous une forme spatialisée sont porteuses d'utilisations multiples et nouvelles pour la connaissance, le suivi et la modélisation des zones érodées. Nous explorons ici des MNT à très haute résolution spatiale, c'est à dire avec des résolutions de quelques centimètres à quelques mètres et avec des qualités de restitution en Z (altitude) meilleures que le mètre. Après une brève présentation des techniques d'acquisition de ces MNT à partir d'images stéréoscopiques ou du Lidar et des précisions qui leur sont associées, on rappelle quelles sont les transformées de l'altitude utiles pour la morphologie. Afin de montrer quelques unes des possibilités qu'ouvrent ces nouvelles données, diverses applications, menées sur les petits bassins versants marneux de l'Observatoire de Recherche et d'Expérimentation de Draix (Alpes, France), sont détaillées. Elles ont trait, d'abord, à l'extraction du réseau des talwegs à partir de MNT à 2 m (et de MNT dégradés à 4, 8, 20 mètres) et au calcul de sa dimension fractale pour les résolutions à 2 m et 8 m ; puis au repérage automatique de ravines sur critères morphologiques simples afin d'aider à la localisation d'aménagements anti-érosifs. On montre enfin les potentialités des topographies précises par Lidar terrestre pour repérer d'étroites rigoles, suivre leur évolution, ou établir des bilans d'érosion sur versant alors qu'une précision d'ordre centimétrique est requise.

In the mountainous and marly catchments of the southern French alps, erosion processes during storm events are particularly intense and the sediment yields are very high. The sediments are transported both as bedload and suspension. In these badlands catchments, the suspended sediment concentration reaches very high values during flood events, frequently over 300 g/l, up to 800 g/l, as measured in the experimental basins of Draix. In this context, the calculation of discharge from the recorded level in the gauging stations is hazardous because the rating curves established with clear water may be no longer appropriated. Therefore, laboratory experiments were conducted in order to determine both the hydrodynamic parameters and rheologic parameters of heavily loaded and hyperconcentrated flows. Based on these parameters it becomes possible to build a reasonable' depth-discharge relationship. The influence of sediment concentration and grain's size distribution on the bulk rheological behaviour of the flow has been evaluated at concentration ranging from 10 to 30 % in weight. The curve fitting for the different concentrations of sediment in suspension is described using the Bingham model. The tilting flume (slope varying from 2 to 6%) experiments show the auto-similarity of the velocity profiles for the different sediment concentrations.We observe, as found in other studies on hyperconcentrated flows, a reduction of average turbulence intensity Tu and an increase of the mean velocity with increasing sediment load. Finally, a standard f (friction factor) vs Re (a suitable Reynolds number) relation has been plotted in order to analyze hydraulic behaviours of fluid over a fixed bed. In addition, these experiments proved the capability of UDV (Ultrasonic Doppler Velocimetry) method to measure the velocities of hyperconcentrated flows.

We estimate agglomeration economies, defined as the effect of density on labour productivity in European regions. The analysis of Ciccone (2002) is extended in two main ways. First, we use dynamic panel estimation techniques (system GMM), thus offering an alternative methodological treatment of the inherent endogeneity problem. Second, the sector dimension in the data allows for disaggregated estimation. Our results confirm the presence of significant agglomeration effects at the aggregate level, with an estimated long-run elasticity of 13 percent. Repeated crosssection regressions suggest that the strength of agglomeration effects has increased over time. At the sector level, the dominant pattern is of cross-sector "urbanisation" economies and own-sector congestion diseconomies. A notable exception is financial services, for which we find strong positive productivity effects from own-sector density.

In order to clarify the erosion processes on a marly bare slope in the Southern Alps, the erosion processes in a steep and erodible slope composed of the Black Marls formation were observed by a time-lapse video camera. The observations revealed that miniature debris flows (MDFs) occurred at the time of the rainfall-runoff event in which the most severe erosion took place in the whole observation period of three months. Analyzing the camera images, the authors showed some characteristics of the MDFs and discussed them in the context of real rainfall-runoff phenomena observed at the outlet of the small experimental basin including the visually observed slope. The following results were obtained. 1) It is roughly estimated that the total amount of sediment discharge by the MDFs was not quantitatively negligible in the total sediment discharge from the entire experimental basin. 2) The MDFs occurred only during the rising limb of the hydrograph during 6 minutes. Since it is considered that the MDFs could transport coarser materials than the normal sediment-laden flow, it is inferred that the grain size distribution of the discharged sediment would temporarily fluctuate greatly. 3) The sediment concentration of the MDFs observed might be largely different from each other. Based on this observation and a review of the literature, in a very steep and highly erodible slope, MDFs or similar phenomena might play an important role in the erosion and transport processes. It is important to consider this possibility when examining the erosion models from a catchment composed of steep and erodible slopes.

In the Southern French Alps, the black marls formation covers a large area and is highly susceptible to weathering and erosion. It has a badlands topography and is subject to high solid transport, bringing high sediment yield downstream and silting up reservoirs. The scale is decisive in the study of erosion processes and quantification of sediment production, de Vente and Poesen (2005). This paper focuses on erosion at the slope and gully spatial scale and at the event temporal scale and is based on the data collected since 1983 by cemagref on a small gully, called Roubine (1330 m²). For the 19852003 period, 1016 rainfall events were registered, 472 produced runoff at the outlet and 373 yielded measurable erosion. A total of 288 sediment trap measurements are available and 205 five floods were sampled for suspended sediment. For the 19 years of the study period, the 19 highest values represent 33% of the total production of the period. For the data of all the events, there was no relation between suspended sediment concentration and flow discharge, but for one event hysteresis curves were observed. The total amount of sediment trap deposit represents a mean 9 m3 per year, i.e. 10 kg y1 m2. The ratio of suspended sediment in the total yield was 15% on average, 20% in cumulated amount, but reached 5050% for a few events. The analysis conducted on 19 years of rainfall-runoff-erosion data and field observations allows us to propose an erosion production model at the gully scale in marly badlands catchments. A seasonal pattern was observed with the renewal of the weathered mantel in winter, substantial displacement of material with spring events, high production of numerous and intense summer storms, and a decrease in sediment availability in autumn.

Located in Alps of south-eastern France, the experimental site of Draix was created in the period 1983-1984 and aim principally at studying mountain erosion processes and corresponding protection devices. Four small watersheds, from 1 000 m2 to 1 km² are equipped to study and quantify runoff and erosion processes, according to the vegetation cover and basin size. The substratum of the basins is black marls, a very erodible ground. As a result, erosion processes and sediment transport are particularly high in that area. The sediment production of a basin is measured at the outlet of the catchments for each storm event: the coarser part of the sediment yield is measured in a sediment trap, the finer part is sampled in the gauging section or monitored continuously with an optical fiber sensor. The sediment concentration during floods may be very high, often over 300 kg m-3, sometimes up to 800 kg m-3. Two prototypes of optical backscattering sensors were designed for these high concentrations in 1993 and have been since used in the field. The optical backscattering sensors are sensitive to the grain size distribution and the sensor calibration must be verified for each flood. The sediment concentration obtained from drying end weighting of samples automatically collected during the event are used to adjust the calibration curve. The relationship between the sensor response and the sediment concentration shows an hysteresis loop for 60% of the floods. This is explained by the evolution of the grain size distribution : the transported sediment may be coarser during the rising limb (unclockwise loop) or during the decay (clockwise loop). Grain size analysis of the samples were realized for several floods with a Malvern particle analyser. The relationship between the different characteristics (grain size, concentration, discharge, rainfall intensity) are pointed out and discussed in terms of erosion and sediment transport processes.

Located in Alps of south-eastern France, the experimental site of Draix was created in the period 1983-1984 and aim principally at studying mountain erosion processes and corresponding protection devices. Four small watersheds, from 1 000 m2 to 1 km² are equipped to study and quantify runoff and erosion processes, according to the vegetation cover and basin size. The substratum of the basins is black marls, a very erodible ground. As a result, erosion processes and sediment transport are particularly high in that area. The sediment production of a basin is measured at the outlet of the catchments for each storm event: the coarser part of the sediment yield is measured in a sediment trap, the finer part is sampled in the gauging section or monitored continuously with an optical fiber sensor. The sediment concentration during floods may be very high, often over 300 kg m-3, sometimes up to 800 kg m-3. Two prototypes of optical backscattering sensors were designed for these high concentrations in 1993 and have been since used in the field. The optical backscattering sensors are sensitive to the grain size distribution and the sensor calibration must be verified for each flood. The sediment concentration obtained from drying end weighting of samples automatically collected during the event are used to adjust the calibration curve. The relationship between the sensor response and the sediment concentration shows an hysteresis loop for 60% of the floods. This is explained by the evolution of the grain size distribution : the transported sediment may be coarser during the rising limb (unclockwise loop) or during the decay (clockwise loop). Grain size analysis of the samples were realized for several floods with a Malvern particle analyser. The relationship between the different characteristics (grain size, concentration, discharge, rainfall intensity) are pointed out and discussed in terms of erosion and sediment transport processes.

In the Southern French Alps, the Black Marls formation, or "Terres Noires" in French, covers a large area. This formation is very susceptible to weathering and erosion. Under the Mediterranean and mountainous climate with frost in winter and high intensity rainfall in summer, it results in "badlands" topography and high solid transport during floods. The experimental site of Draix was created in the period 1983-1984 and aim principally at studying erosion processes and flood generation in this area. Four small watersheds, from 1 000 m2 to 1 km² are equipped to study and quantify runoff and erosion processes, according to the vegetation cover and basin size. The analyses of the data at the event scale pointed out the non-linearity of the response to a rainfall input and revealed the role of the deposition and scouring processes in the channel network. In order to study how this phenomenon control the erosion response at the outlet, field observations were carried out throughout the year in the 2 basins devoid of vegetation : Laval (0.86 km²) and the Moulin (0.09 km²). The paper analyses the field observations and collates it to the rainfall inputs and to the runoff and sediment supplies at the outlet for the same rainfall events or group of events. In addition, the water and sediment response of a monitored gully (The Roubine, 1330 m²) are used to evaluate the contribution of the slopes for the same events. The analysis of the sequence shows how the spring and summer rainfall events increase the temporary stocks in the gully bottoms and channel reaches even when the sediment output is high. At the beginning of autumn, the upstream part of the network begins to empty and the high floods of November scoured most of the remaining material in the larger channels. These results allow to build a seasonal model of the erosion and transport processes in the marly badlands basins.

In order to clarify the erosion processes on a marly bare slope in the Southern Alps, the erosion processes in a steep and erodible slope composed of the Black Marls formation were observed by a time-lapse video camera. The observations revealed that miniature debris flows (MDFs) occurred at the time of the rainfall-runoff event in which the most severe erosion took place in the whole observation period of three months. Analyzing the camera images, the authors showed some characteristics of the MDFs and discussed them in the context of real rainfall-runoff phenomena observed at the outlet of the small experimental basin including the visually observed slope. The following results were obtained. 1) It is roughly estimated that the total amount of sediment discharge by the MDFs was not quantitatively negligible in the total sediment discharge from the entire experimental basin. 2) The MDFs occurred only during the rising limb of the hydrograph during 6 minutes. Since it is considered that the MDFs could transport coarser materials than the normal sediment-laden flow, it is inferred that the grain size distribution of the discharged sediment would temporarily fluctuate greatly. 3) The sediment concentration of the MDFs observed might be largely different from each other. Based on this observation and a review of the literature, in a very steep and highly erodible slope, MDFs or similar phenomena might play an important role in the erosion and transport processes. It is important to consider this possibility when examining the erosion models from a catchment composed of steep and erodible slopes.

(trad auto)The objective of the work is to determine the solid and liquid flows produced by a small mountain watershed in response to a rainy event. It is based on observations made in the Draix research basins, studies the processes and factors involved in the formation of flows and quantifies their respective roles. Different spatial scales of analysis are used: the plot, the elementary watershed or gully and the watershed. The strong non-linearity of the responses is highlighted by an in-depth statistical analysis of the available data (1985-2003). This statistical work is supplemented by the interpretation of specific observations and field experiments: rain simulation experiments and observations under natural rainfall at the plot and ravine scale, monitoring the state of sediment stocks in the beds. Finally, a first spatialized modeling draft, using the ETC (Erosion of torrents in flood) model developed at Cemagref, was carried out, verifying the model's ability to reproduce the observed temporal and spatial functioning.; Le travail a pour objectif de déterminer quels sont les flux solides et liquides produit par un petit bassin versant de montagne en réponse à un épisode pluvieux. Il s'appuie sur des observations réalisées sur les bassins de recherche de Draix, étudie les processus et facteurs intervenant dans la formation des écoulements et quantifie leurs rôles respectifs. Différentes échelles spatiales d'analyse sont utilisées : la parcelle, le versant ou la ravine élémentaire et le bassin versant. La forte non-linéarité des réponses est mise en évidence au moyen d'une analyse statistique approfondie des données disponibles (1985-2003). Ce travail statistique est complété par l'interprétation d'observations et d'expérimentations de terrain spécifiques : expérimentations de simulation de pluie et observations sous pluies naturelles à l'échelle de la parcelle et de la ravine, suivi de l'état des stocks de sédiments dans les lits. Enfin, une première ébauche de modélisation spatialisée, à l'aide du modèle ETC (Erosion des torrents en crue) développé au Cemagref, a été réalisée, vérifiant la capacité du modèle à reproduire le fonctionnement temporel et spatial observé.

These tests emphasize the great potential of this type of vector to produce low-cost images with a very high resolution. This study demonstrates that these products can be used in rough landscapes and areas that are hard to reach such as badlands. They are well adapted for high-scale and highly repetitive mapping on small experimental catchments such as those found at the Draix observatory. They are of great interest in the study of erosion issues because they provide a link between the local-scale survey (ground data) and the traditional remote sensing survey (aerial photograph or satellite). Such products can thus help further the understanding of hydrological measurements at the catchment's outlet.

Trois ensembles de bassins expérimentaux permettent d'étudier, en France méditerranéenne, les conséquences de la forêt sur les crues. Les résultats différent de façon considérable d'un bassin à l'autre, montrant la complexité des relations pluies-débits. Il est à noter que l'augmentation des crues extrêmes est loin d'être avérée dans tous les cas. Une réflexion sur le fonctionnement hydrologique des bassins permet de lever la contradiction apparente entre les résultats observés.

National audience; Cemagref (Agricultural and Environmental Engineering Research Center) in Grenoble and the Scientific Group "Draix, mountain erosion studies" held a conference in Digne, in the French Alps, from October 15 to 17, 2003, devoted to the theme of gully erosion in mountain areas. The aims of this international seminar were to evaluate the state of the art in research on erosion by water in uplands, to allow scientists to meet and exchange ideas on related topics, and to initiate further collaborations, with particular emphasis on field work in research basins. This meeting, which encompassed 32 oral presentations and 24 posters, was attended by 80 participants from 12 countries. A one-day scientific excursion to the Draix experimental site was dedicated to multidisciplinary researches on mountain erosion and upland floods. Gully erosion represents an important sediment source in river systems and accounts for as much as 70 to 90% of the overall sediment production of a catchment. In mountain areas, the steep slopes enhance gully processes, accelerate sediment transfer from uplands to valley bottoms and generate natural hazards: mudflows, overflowing of heavily loaded floods, silting up of reservoirs, for example. Therefore, there is a need for monitoring and for experimental and modelling studies in order to forecast the effects of climatic or land use changes on sediment dynamics in mountain catchments. In mountain environments, the development of appropriate measurement and observation techniques is quite difficult. Original devices were presented, operating from the small plot scale to the large gully and catchment scales. Several presentations focused on the processes and factors controlling erosion and gullying in mountain areas: these include mechanical breakdown and chemical weathering of rocks, bed load and suspension sediment transport, in which threshold effects have a determining influence. Mass movements are also a substantial source of sediments in steep upland basins. These may occur as superficial earth flows affecting the weathered layer, and as deep-seated landslides or rock-block slides. Several papers examined the susceptibility of the terrain to the mass movements, the conditions for landslide initiation, the relations between mass movements in black marls and rockfalls in the overlying limestone levels, and the formation and propagation of debris flows. The major role of a few paroxystic events in sediment production was pointed out by several authors. Particle detachment, mass movement initiation, and sediment transport are processes subjected to threshold effects. The connectivity in the basin network may vary from one event to the other. The studies on sediment load rates during floods showed that the availability and location of sediment stocks in the channel and gully bottoms play a greater role than the transport capacity of the flow. The role of the vegetation cover in diminishing sediment yields is widely accepted. Therefore, it is still necessary to understand how the aerial parts and roots of vegetation interact with runoff and erosion in order to model the evolution of gully systems subject to climatic or land use changes. These studies show that erosion modelling becomes more complex as one goes from the plot or slope scale to the catchment scale, which necessitates a transport-distance approach. At this scale, the question of connectivity between flow paths becomes crucial and depends on the time scale of the study. The present issue of the journal publishes five papers related to erosion measurement and remote sensing, to erosion dynamics at the catchment or at a regional scale, and interactions between erosion, and vegetation and restoration works. The paper by D. Raclot et al. presents a first attempt of at using low-altitude aerial photographs taken from a drone for erosion surveys in the Draix area. The relatively poor vertical accuracy (70 cm) does not allow yet for the measurement of short-term ablation, but the good quality and resolution of the pictures, and the possibility for of repeated acquisition offer good opportunities for the survey of landform evolution, especially in remote areas of the studied catchment. Two others additional papers (E. Gauché, J. Collinet & P. Zante) present fruitful methodologies for mapping and topographic survey of gully development. Recent progress in GIS and DEM techniques complete the quantitative observation of erosion processes. Addressing scale issues, the paper by J. Collinet and P. Zante presents different techniques in Tunisia covering rainfall simulation on plots in the flat upstream areas, monitoring of gully evolution on the intermediate slopes, sedimentation in reservoirs at the outlets. The authors show the importance of the processes linking these three slope interfaces. In Morocco, E. Gauché found contrasting rhythms of landscape evolution between mountain slopes and alluvial deposits downstream. Lithology and relief play a role, which is reinforced by land use change to a point that an erosion crisis may be identified, mainly due to the decrease of the extent of tilled areas. The paper by F. Rey presents an experimental bioengineering device for trapping sediments in gullies in marls of the French Alps and shows how brush layers 1.2 m thick across a gully may trap 0.1 to 0.3 m3 per year. The paper by M. Cohen and F. Rey discusses the paradox between positive vegetation dynamics at the regional scale and active erosion observed at the local scale and shows that processes vary depending on spatial scales: for instance, gully erosion may co-exist with vegetation spread on interfluves. Temporal and spatial scales as well as connectivity in soil erosion systems appear as key questions for further research on gully erosion, including appropriate measurement techniques, process analysis, modelling, and upscaling.

[Departement_IRSTEA]RE [TR1_IRSTEA]RIE / TRANSFEAU; National audience; Research on the effects of forest on river flooding in Frances's mediterranean regions has been conducted based on experiments in three catchment areas. Results display considerable differences from one area to another, highlighting the complexity of the nexus rainfall-flowrates. It should be noted that an increase in extreme flooding is far from being clear in every case. Reflecting on overall hydrological behaviour in the catchment areas has enabled researchers to explain the apparent contradiction in the results obtained.; Trois ensembles de bassins expérimentaux permettent d'étudier, en France méditerranéenne, les conséquences de la forêt sur les crues. Les résultats différents de façon considérable d'un bassin à l'autre, montrant la complexité des relations pluies-débits. Il est à noter que l'augmentation des crues extrêmes est loin d'être avérée dans tous les cas. Une réflexion sur le fonctionnement hydrologique des bassins permet de lever la contradiction apparente entre les résultats observés.

In the southern French Alps, the black marl formation covers a large area. It is very sesitive to weathering and erosion and presents a characteristics badlands morphology with V-shaped gullies and high levels of solid transport during floods. Many studies on similar badlands basins have demonstrated the variability of this terrain's erosional response to a rainfall event. This work identifies four types of rainfall-erosion situations.

Autres; Ce site Web présente ORE Draix qui est un Observatoire de Recherche en Environnement consacré à l'étude des processus hydrologiques et érosifs en montagne. L'Observatoire bénéficie du soutien du Ministère Délégué à la Recherche et aux Nouvelles Technologies.