Search

Nitrous oxide (N2O) is a greenhouse gas almost 300 times more powerful than CO2 in terms of global warming potential, and it is also the first ozone-depleting substance emitted in the 21st century. Approximately 43% of N2O emissions are estimated to be due to anthropogenic activities worldwide, and 52% of this anthropogenic part come from cultivated soils. The main cause of anthropogenic emissions is nitrogen fertilization.The production, transfer and emission of N2O from soils are complex multifactorial processes, with a high spatial and temporal variability. Although N2O production in soils has multiple origins, the main source remains denitrification reactions during microbial respiration under anaerobic conditions. Thus, one of the major soil control factors is the availability of oxygen to soil organisms, which partly depends on the soil structure. The spatiotemporal variability of N2O emissions is explored by deterministic studies that focus either on the soil microstructure scale, i.e. the scale of N2O production and microorganism habitat, or on the macrostructure scale, to focus on fluids transfers. However, the influence of soil micro- and macrostructure studied together on N2O emissions is still poorly known, and represents the objective of this work.A multi-scale approach was adopted to better understand the determinism of N2O emissions. The spatial variability of N2O emissions at the field scale was estimated during a snap-shot campaign on the same soil type with contrasted structural states, induced by different agricultural practices (4 soil modalities crossing strip-till and tillage with compacted or uncompacted areas). 24 soil cylinders were collected in low and high N2O emission zones and were then scanned by using both X-ray macro- and micro-tomography. Quantitative morphological tools were used to describe soil structure at the macro and micro scales while simultaneously studying other soil properties influencing N2O emissions (air permeability, gas diffusivity, nitrogen, pH, soil texture, etc.). The 4 soil modalities studied showed contrasted N2O emissions along with contrasting macrostructural and gas transfer indices. The ongoing work is aimed at clarifying the relationships between multiscale soil structure, gas transfer and other soil factors on N2O emissions.

Sustainable crop production implies high efficiency of field operations and protection of the soil as a natural resource. Soil physical fertility is threatened by compaction, especially deep soil horizons for which remediation is more critical. Increased soil compaction, linked to the increase of agricultural equipment weight, causes yield losses on spring and summer crops. To avoid soil compaction and ensure field operations efficiency, including satisfactory crop production in a cost-effective way, field readiness prediction is necessary. Field readiness is defined by the combination of soil workability (soil suitability for cultural operations) and soil trafficability (soil capacity to support machinery during traffic without soil physical degradation). Available tools focused on one part of the problem, e.g. soil compaction risk in deep soil horizon, or possibility of efficient field operation ; each has usually been built for a specific pedoclimatic context, which questions its application in a broader context. Available tools also frequently need to be upgraded to better consider compaction risk according to machinery evolution. The J-DISTAS project (2019-2022) aims at evaluating and improving these tools, and structuring them to create a prototype of interoperable tools to predict field readiness. The resulting tool will be based on the combination of two mechanistic models (Terranimo for soil compaction and the CHN crop model for soil water content), pedotransfer functions to estimate soil water potential and soil workability, and a decisional tool of field readiness build from expert knowledge. Its ability to predict field readiness and its sensibility to input data will be evaluated. The developed inter-operable tools could be used as a decision support tool that includes field readiness in strategic decisions, conception of cropping systems in the context of global changes, or optimization of mechanical cost for equipment in agricultural machinery, and will help to soil physical quality protection.

N2O emission in soils is a consequence of the activity of nitrifying and denitrifying microorganisms and potentially abiotic processes. However, the large microscale variability of the soil characteristics that influence these processes and in particular the location of anoxic microsites, limits prediction efforts. Better understanding of denitrification activity on microscopic scales is required to improve predictions of N2O emissions.This study explored the role of soil microstructure on N2O emission. To fulfill this objective we sampled 24 soil columns (5 cm diameter, 6 cm height) in the surface layer of a same plot in a cultivated soil (Luvisol, La Cage, Versailles, France). The soil samples were saturated with a solution of ammonium nitrate (NH4NO3), and equilibrated at a matrix potential of -32 cm (pF 1.5). The emitted fluxes of N2O were measured during 7 days. At the end of the experiment, the soil columns were scanned in a X-ray micro tomograph, at the University of Poitiers. A 32 µm voxel resolution was achieved for the 3D reconstructed images.In order to reduce noise and segment the 3D images, the same protocol was implemented for all columns. The reduction of noise consisted of passing a non-local mean filter, a non-sharp mask and a radial correction. Such combination of steps succeeded in removing both ring artifacts and the radial dependence of the voxel values. Due to the variety of material densities in the soil, a local segmentation based on the watershed method was implemented to classify the soil constituents in four classes (based on its density value): air, water and organic matter (OM), soil matrix and minerals. This method is good for detecting thin pores and avoids missclassification of voxels undergoing partial volume effect, which can lead to false organic coatings around macropores.The soil columns exhibited a large variability of accumulated N2O after 7 days (from 107 to 1940 µgN kg-1 d.w. soil). The size of OM clusters varied between a couple and up to thousands of voxels. No correlation was found between the emission of N2O and the porosity, nor between the N2O emission and the connectivity of the air phase. Based on the premise that the less accessible is the oxygen to the OM, the bigger should be the N2O emission of the soil column, we proposed and computed a microscopic spatial descriptor, Igd, based on the notion of the geodesic distance between clusters of OM and air for each soil column 3D image. We expect to find a correlation between Igd and the N2O emission.

The structure of soils, i.e. the macroscopic organization of aggregates and pores, conditions the storage and transport of water and gas in the soil, and strongly determines the physico-chemical environment of soil organisms (plants, micro and macro-organisms). The description of the soils structure dynamics constitutes a major issue in the current context of global change, at the scientific, environmental and agronomic level. However, few tools are available to monitor this dynamic non-destructively and in situ. We therefore propose to develop a new method based on the analysis of acoustic emissions (AE) spontaneously emitted by soils during the evolution of their structure. A laboratory feasibility study was conducted to explore the links between variations in soil structure and the AE emitted during soil desiccation.Two undisturbed soil columns (8 cm in diameter, 5 cm high) were sampled in an agricultural field (near Chartres in France), in the surface horizon of a Glossic Retisol. These cylinders were air dried (20°C during 9 days), and the AE produced during drying were monitored using piezoelectric sensors place at the soil surface. The concomitant soil structure changes were followed through 3D images, acquired by X-ray tomography (CIRE platform, INRAE, Nouzilly) all along the experiment. These images, with a resolution of 168 µm, were used to characterize the pore network (porosity, surface density, connectivity, etc.).The dynamics of the EAs recorded during the drying of the samples is comparable for the two samples: the AE rates are maximum at the start of the experiment and then reach a plateau. Changes in soil structure follow the same dynamics, e.g. considering porosity or surface density of the pores. If we analyze the relationship between the signals recorded by the surface sensors (EA rate) and the porosity, we observe a linear relationship (R² of 0.79). This relationship, although encouraging, remains to be consolidated by additional results.To go further, it is also necessary to define the necessary conditions to perform such a measurement in situ, and to improve the acoustic signal processing to characterize the EA produced during soil desiccation. Indeed, a major objective of our work is to differentiate, thanks to EAs, the various factors responsible for the evolution of soil structure (physical and biological), by determining their "acoustic signature".

The measurement of aggregate stability is an important indicator of soil quality and is widely used for monitoring soil condition. The SLAKES mobile app is an alternative tool to laboratory-based methods to measure soil aggregate stability. It provides aggregate stability measurements through Slaking Index (SI) with SI close to 0 suggesting high stability and values above 7 suggesting minimum stability. As the duration of this low-cost experiment is only 10 minutes, SLAKES is very attractive for scientists and no-scientists. SLAKES was implemented in Australia and has proven its efficiency in several studies.This study was conducted to determine whether the SLAKES mobile app could be adapted to French soils and then could be an alternative to the Mean Weight Diameter (MWD) method, normalized in France (ISO 10390). More specifically, the three main objectives were: (i) determining whether the aggregate stability measurements depend on the phones used for the experiment, (ii) estimating the number of measurements necessary to get reliable results, (iii) determining whether the app has the ability to detect the effect of contrasting soil managements previously shown using the MWD method.The study was performed on silty loam soils from EFELE (Effluents d’Elevage et Environnement) experimental site at le Rheu (Brittany, France) which is part of the French “Organic Residues” research observatory (SOERE PRO). The experimental design combines two different tillage practices (conventional tillage and shallow tillage) with two fertilizer treatments (mineral and organic (cattle manure)) randomly replicated three times. Soil samples were collected in March 2017 at both 0-15 cm and 15-25 cm depth from the 12 plots. The SI was measured on three aggregates simultaneously and this measurement was repeated 15 times for each sample which provided 45 SI per sample. Outliers above SI=11 were removed before statistical treatments. Four different phones of the same brand and generation were used to measure SI.An analysis of variance showed that the effect of the smartphone on SI measurements was not significant (p-value = 0.73, 0.88, 0.067 for 3 different samples). The SLAKES results showed comparable significant separation of means (p < 0.0001) between each soil management than the MWD method. According to the results of an analysis of variance, SI was significantly lower in reduced tillage than in conventional tillage condition (p-value = 2.10-16). These results indicate a higher soil stability in reduced tillage. Yet, the analysis of variance did not underline any effect of fertilization on the SI (p-value = 0.28), as previously found with the MWD method.This study proved that a relatively simple mobile app can detect the effect of soil management practices on aggregate stability with a similar performance than the MWD method. This conclusion was reinforced by the existing correlation between the SI and the MWD index (p-value = 0.00059, R2 = 0.39). We recommend to perform similar experiments on other sampling campaigns or in other pedological and soil management contexts taking at least 15 measures per sample.

International audience; During the last 10 years, several attempts to map soil attributes at the scale of mainland France have been realised. We exemplify them by seven major outputs: maps of organic C stocks, trace elements (TE), microbial density and diversity, soil thickness, available water capacity (AWC), extractable P, and changes in soil pH. We first briefly describe the data and the methods used to produce these maps and summarise their main results. We then focus on their impacts on various categories of the public, i.e. the general public and citizens; farmers; private companies; non-governmental organisations; agricultural development organisations, stakeholders, and national agencies; French governmental bodies; and international organisations. We also analyse the demands that came to the French National Soil Information Centre from 2008 to 2018 and the impact that our activities had in various media. Soil organic C had the largest impact in nearly all categories of end-users, which may be linked to the recent ‘4 per 1000’ initiative launched by the French governmentduring the COP21 and to the fact that farmers are interested in increasing the organic matter content of their soil for increasing the fertility. TE obtained high scores, which may be related to citizens' care about health and to the fact that governmental bodies and national agencies have a major interest in site contamination assessments. The soil P content, pH, and AWC exhibited major impacts on the agricultural sector. Maps of the soil P content and pH were used as geomarketing tools by private companies selling fertilisers and soil amendments, whereas the AWC was already incorporated into decision-making aid tools for irrigation management developed by development organisations for farmers. Microbial diversity generated collaborations with a large network of farmers and had a large media impact. Nevertheless, the visibility of soil information to the general public should be increased. This can be done by using new multimedia and interactive tools. Overall, these selected examples of digital soil mapping of soil attributes at the national scale in France clearly indicate that the soil attributes have substantial impact on various categories of end-users, such as farmers, professional organisations, stakeholders, and policymakers at different levels of decision-making, among others. However, the impacts on the general public and citizens are more difficult to quantify, and increasing the soil awareness of the general public should be of high priority.

Agroforestry systems are promoted for providing a number of ecosystem services and environmental benefits, including soil protection and carbon sequestration. This study proposes a modelling approach to quantify the impact of soil redistribution on soil organic carbon (SOC) storage in a temperate hedgerow landscape. Evolution of SOC stocks at the landscape scale was examined by simulating vertical and horizontal SOC transfers in the 0–105 cm soil layer due to soil redistribution by tillage and water processes. A spatially explicit SOC dynamics model (adapted from RothC-26.3) was used, coupled with a soil-redistribution model (LandSoil). SOC dynamics were simulated over 90 years in an agricultural hedgerow landscape dedicated to dairy farming, with a mix of cropping and grasslands. Climate and land use were simulated considering business-as-usual scenarios derived from existing information on the study area. A net decrease in SOC stocks was predicted at the end of the simulation period. Soil redistribution induced a net SOC loss equivalent to 2 kg C ha−1 yr−1 because of soil exportation out of the study site and an increase in SOC mineralization. Hedgerows and woods were the only land use in which soil redistribution induced net SOC storage. Soil tillage was the main process that induced soil redistribution within cultivated fields. Soil exportation out of the study area was due to erosion by water, but remained low because of the protective role of the hedgerow network. These soil transfers redistributed SOC stocks in the landscape, mostly within cultivated fields.

International audience; This work presents the first high-resolution map of soil organic carbon (SOC) in mainland France, including soils below 30 cm. The research was performed within the framework of GlobalSoilMap (GSM). SOC predictions for different depth layers (0–5 cm, 5–15 cm, 15–30 cm, 30–60 cm, 60–100 cm and > 100 cm) were made at 90 and 500 m resolution for mainland France, along with their upper and lower confidence intervals. The maps were developed using data mining and an elaborate cross-validation scheme. The 90 m maps were compared to 500 m resolution GlobalSoilMap maps and the SoilGrids1km (SG1km) product. The latter is a global model for predicting soil properties for the same depth layers, at 1 km resolution. At 90 and 500 m resolution, the predicted SOC content was unbiased and showed good agreement with the measured SOC, despite the poor model diagnostics and decrease of performance with depth. It was found that the subsoil (> 30 cm) carbon pool for France contributes 49% to the total soil carbon stock. The use of coarser resolution prediction grids resulted in smoother spatial patterns and wider confidence intervals; however, it did not bias the estimated carbon stocks. Applying GlobalSoilMap specifications to France, using a large soil dataset and all the exhaustive spatially available data outperformed SG1km predictions. The overall spatial patterns of the SG1km SOC content were found to be very similar to the GlobalSoilMap maps. However, the SG1km overestimated the SOC content and carbon stocks (> 75% for the total carbon stock, and 100% for the stocks below 30 cm) and showed a similar spatial distribution over the different soil depth layers. The main reason for the overestimation was that the local data used in SG1km was rather small (56 samples) and not representative in terms of SOC content or represented soil types; the profiles had far higher SOC content and this may have propagated in the modelled vertical profile and the kriging part of the residuals. Improvements for SG1km may entail the use of a representative national subsample from large national soil databases. Furthermore, a bottom-up approach such as GlobalSoilMap may be more favourable when considering prediction accuracies, data privacy policies and local acceptance of generated products.

ISI Document Delivery No.: HT3RS Times Cited: 0 Cited Reference Count: 79 Chen, Songchao Mulder, Vera Leatitia Martin, Manuel P. Walter, Christian Lacost, Marine Richer-de-Forges, Anne C. Saby, Nicolas P. A. Loiseau, Thomas Hu, Bifeng Arrouays, Dominique French Ministry of Ecology; French Environment and Energy Management Agency (ADEME); French Institute for Research and Development (IRD); French Institute for National Geographic and Forest Information (IGN); National Institute for Agronomic Research (INRA); China Scholarship Council [201606320211]; French Scientific Group of Interest on soils (GIS Sol) The collection and handling of soil data was supported by the French Scientific Group of Interest on soils (GIS Sol), including the French Ministry of Ecology, the French Ministry of Agriculture, the French Environment and Energy Management Agency (ADEME), the French Institute for Research and Development (IRD), the French Institute for National Geographic and Forest Information (IGN) and the National Institute for Agronomic Research (INRA). We thank all the people involved in sampling the sites and populating the database. We would also like to thank Quentin Styc and Philippe Lagacherie for sharing their ideas on dealing with right censored data. Songchao Chen received the support of the China Scholarship Council for three years' of Ph.D. study in INRA and Agrocampus Ouest (under grant agreement no. 201606320211). 0 Elsevier science bv Amsterdam 1872-6259; International audience; Soil thickness (ST) is a crucial factor in earth surface modelling and soil storage capacity calculations (e.g., available water capacity and carbon stocks). However, the observed depths recorded in soil information systems for some profiles are often less than the actual ST (i.e., right censored data). The use of such data will negatively affect model and map accuracy, yet few studies have been done to resolve this issue or propose methods to correct for right censored data. Therefore, this work demonstrates how right censored data can be accounted for in the ST modelling of mainland France. We propose the use of Random Survival Forest (RSF) for ST probability mapping within a Digital Soil Mapping framework. Among 2109 sites of the French Soil Monitoring Network, 1089 observed STs were defined as being right censored. Using RSF, the probability of exceeding a given depth was modelled using freely available spatial data representing the main soil-forming factors. Subsequently, the models were extrapolated to the full spatial extent of mainland France. As examples, we produced maps showing the probability of exceeding the thickness of each GlobalSoilMap standard depth: 5, 15, 30, 60, 100, and 200 cm. In addition, a bootstrapping approach was used to assess the 90% confidence intervals. Our results showed that RSF was able to correct for right censored data entries occurring within a given dataset. RSF was more reliable for thin (0.3 m) and thick soils (1 to 2 m), as they performed better (overall accuracy from 0.793 to 0.989) than soils with a thickness between 0.3 and 1 m. This study provides a new approach for modelling right censored soil information. Moreover, RSF can produce probability maps at any depth less than the maximum depth of the calibration data, which is of great value for designing additional sampling campaigns and decision making in geotechnical engineering.

We report observations of acoustic emissions (AE) from growing plant roots and burrowing earthworms in soil, as a noninvasive method for monitoring biophysical processes that modify soil structure. AE emanating from earthworm and plants root activity were linked with time-lapse imaging in glass cells. Acoustic waveguides where installed in soil columns to monitor root growth in real time (mimicking field application). The cumulative AE events were in correlation with earthworm burrow lengths and with root growth. The number of AE events recorded from the soil columns with growing maize roots were several orders of magnitude larger than AE emanating from bare soil under similar conditions. The results suggest that AE monitoring may offer a window into largely unobservable dynamics of soil biomechanical processes such as root growth or patterns of earthworm activity - both important soil structure forming processes.

National audience

International audience

International audience; In this work we aimed at developing a conceptual framework in which we improve our understanding of the controlling factors for soil organic carbon (SOC) over vast areas at different depths. We postulated that variability in SOC levels may be better explained by modeling SOC within soil-landscape systems (SLSs). The study was performed in mainland France, and explanatory SOC models were developed for the sampled topsoil (0–30 cm) and subsoil (>30 cm), using both directed and undirected data-mining techniques. With this study we demonstrated that there is a shift in controlling factors both in space and depth which were mainly related to (1) typical SLS characteristics and (2) human-induced changes to SLSs. The controlling factors in relation to depth alter from predominantly biotic to more abiotic with increasing depth. Especially, water availability, soil texture, and physical protection control deeper stored SOC. In SLSs with similar SOC levels, different combinations of physical protection, the input of organic matter, and climatic conditions largely determined the SOC level. The SLS approach provided the means to partition the data into data sets that were having homogenous conditions with respect to this combination of controlling factors. This information may provide important information on the carbon storage and sequestration potential of a soil.

Summary Soil organic carbon (SOC) sequestration and soil redistribution are linked to soil properties, land use, farming system and climate. In a global-change context, landscape and climate changes are expected and will most probably have impacts on changes in the soil. Soil change was simulated from 2010 to 2100 in an 86-ha hedgerow landscape under different scenarios of landscape and climate changes. These scenarios combined contrasting land uses, hedgerow networks and climates. Two models were combined to evaluate the impact of these scenarios on soils: LandSoil, a soil redistribution model, and a SOC model based on RothC-26.3. A soil thickness of 105 cm was considered. The results indicate that the main factor influencing soil degradation was land-use change: when compared with the baseline business-as-usual landscape, the landscape with the most intensive agricultural systems had the greatest soil erosion (+0.26 t ha−1 year−1) and reduced mean SOC stocks (−17 t ha−1 after 90 years). The second significant factor was climate change, followed by hedgerow network density. Sensitivity to climate change differed between landscapes, and the most sensitive were those with continuous winter wheat. The results indicate that a hedgerow landscape is well adapted to protect soil (regarding carbon storage and soil erosion) in a context of climate change. However, this type of landscape is highly sensitive to cropping intensification and should be protected.

International audience; Tile drainage may have contrasting effects on soil nitrous oxide (N2O) emission. Because drainage decreases anoxic periods in soils, it could reduce N2O production via denitrification and also limit the reduction of N2O into nitrogen gas (N2). Moreover, drainage accelerates the discharge of water enriched in dissolved N2O and mineral nitrogen. Thus, nitrogen losses and N2O releases from discharged surface water need to be quantified to assess the total effect of drainage on N2O emissions. Thus, the objectives of this study were two-fold: (1) to assess the effect of tile-drainage on soil N2O emissions in an agricultural area in Central France (direct emissions) and (2) to compare emissions from soils and from the stream draining the area (indirect emissions). The emissions of N2O by soils were measured using static chambers in two drained and two undrained cereal plots over two growing seasons. A rule-based model was fitted to identify the influence of drainage and ancillary variables. Stream N2O emissions were measured with a floating chamber during one growing season. The mean direct N2O emissions were 0.071 mg N m−2 h−1 and were larger in the undrained plots than in the drained plots in both growing seasons (p < 0.001). The rule-based model showed that the drainage effect on N2O emissions was dominant over the permanent soil variables. The mean stream N2O emissions were 0.190 mg N m−2 h−1. The surface water emissions represented 31 kg N during the flow period (7 months) while direct emissions were 1846 kg N during the same period. Thus, indirect emissions accounted for

The objective of this work was to compare three spatially explicit modeling approaches for soil depth (SD) for France (540 K km 2 ), produced using: i) a straight forward digital soil mapping (DSM) approach, based on regression treemodeling ( RTM ), ii) gradient boosting modeling ( GBM ), and iii) multi-resolution kriging ( MrK ) for large datasets. SD was defined from the USDA Soil Survey Manual as the depth (in cm) to a lithic or paralithic contact. SD was determined for 2116 sites from the French Soil Monitoring network (RMQS) and 14,113 sites from French Soil Inventory (IGCS) program. The RMQS dataset had a better spatial coverage and was collected following a standardized procedure. The RMQS dataset was used for calibration; the IGCS dataset was used for validation. SD ranged from 0 to 300 cm for RMQS (mean = 102 cm), and from 3 to 295 cm for IGCS (mean = 93 cm). Exhaustive environmental data were used to characterize the climate, the organisms, the topography and the other known soil forming factors, following the Soil-Landscape paradigm. The final maps were compared using the same strategy; the maps' accuracy and difference were assessed by comparing the SD predictions to observed data (RMQS and IGCS datasets) and to an available soil map (1:1M). The three DSM approaches predicted the SD trend mainly from the covariates derived from the digital elevation model (DEM). Then, the most important covariates were soil properties, climate covariates, and finally land use. The three predictive maps showed similar accuracy, and were consistent with the 1:1M SD map. The four maps presented similar spatial pattern at the country scale, but the RTM and GBM map showed the higher spatial heterogeneity, while the MrK map was smoother. The three maps had poor performance to estimate the shallower and deeper SD values. This issue was discussed and three options were presented to solve it (implementing calibration dataset, addition of kriged residuals, quantile transformation). Finally, this work showed that the choice of the modeling approach should be done considering the users' goals.

International audience

International audience

Water in soil is known to be a key factor for controlling N2O emissions because N2O is mainly produced by denitrification in anoxic environments. In this study, we proposed a methodology to image the water and soil structure of a soil sample with X-ray computed tomography while controlling the hydric state and monitoring N2O fluxes. We used a multistep outflow system to apply two wetting–drying cycles to an undisturbed soil. The soil core was scanned with coarse-resolution X-ray computed tomography, one time during wetting and several times during drying, to measure quantitative and qualitative indicators of the pore network. Nitrous oxide emissions were higher during the first (C1) than during the second (C2) wetting–drying cycle for both the wetting and the drying phases. Fluxes increased quickly after the beginning of the drying phase to reach a peak after 5 h. Differences in the intensity of N2O emissions between the two cycles were attributed to differences in the water saturation, air-phase connectivity, and relative gas diffusion coefficient, which led to more or less N2O production, consumption, and entrapment in the soil. The speed of the N2O emissions at the beginning of the drying phase depended on the rate of increase of the air-filled pore volume and connectivity, and was especially well described by the estimated relative gas diffusion coefficient. Parameters of the soil structure were not able to explain completely the intensity of N2O emissions during drying; N2O production and consumption factors were also involved.

International audience

Soil organic carbon (SOC) has a crucial impact on global carbon storage at world scale. SOC spatial variability is controlled by the landscape patterns resulting from the continuous interactions between the physical environment and the society. Natural and anthropogenic processes occurring and interplaying at the landscape scale, such as soil redistribution in the lateral and vertical dimensions by tillage and water erosion processes or spatial differentiation of land-use and land-management practices, strongly affect SOC dynamics. Inventories of SOC stocks, reflecting their spatial distribution, are thus key elements to develop relevant management strategies to improving carbon sequestration and mitigating climate change and soil degradation. This study aims to quantify SOC stocks and their spatial distribution in a 1,000-ha agricultural bocage landscape with dairy production as dominant farming system (Zone Atelier Armorique, LTER Europe, NW France). The site is characterized by high heterogeneity on short distance due to a high diversity of soils with varying waterlogging, soil parent material, topography, land-use and hedgerow density. SOC content and stocks were measured up to 105-cm depth in 200 sampling locations selected using conditioned Latin hypercube sampling. Additive sampling was designed to specifically explore SOC distribution near to hedges: 112 points were sampled at fixed distance on 14 transects perpendicular from hedges. We illustrate the heterogeneity of spatial and vertical distribution of SOC stocks at landscape scale, and quantify SOC stocks in the various landscape components. Using multivariate statistics, we discuss the variability and co-variability of existing spatial organization of cropping systems, environmental factors, and SOM stocks, over landscape. Ultimately, our results may contribute to improving regional or national digital soil mapping approaches, by considering the distribution of SOC stocks within each modeling unit and by accounting for the impact of sensitive ecosystems.

Present address for the author M. Lacoste : INRA, UR 0272 Science du sol, Centre de recherched'Orléans, CS 40001 Ardon, 45075 Orléans Cedex 2; Erosion is one of the main threats to soils and is associated with numerous environmental and economic impacts. At the landscape scale, soil redistribution patterns induced by water and tillage erosion are complex, and landscape structures play an important role in their spatial distribution. In this study, soil redistribution patterns, generated by both water and tillage erosion, were estimated in the vicinity of hedges in an agricultural landscape. Two complementary methods were employed to estimate soil redistribution from 1960 to 2010: 137Cs conversion models and a spatially distributed soil erosion model (LandSoil model). Both methods determined that hedges affected soil redistribution patterns, which led to soil deposition or limited soil erosion uphill from hedges, even though soil erosion rates were consistently higher than soil deposition rates. Depending on the method, mean soil redistribution rates ranged from − 15.9 to − 4.7 t ha− 1 yr− 1 for all sampling points, from − 4.8 to 2.2 t ha− 1 yr− 1 in positions uphill from hedges and from − 4.8 to − 11.2 t ha− 1 yr− 1 in positions located downhill from hedges. The impact of tillage on soil redistribution in the vicinity of hedges was found to be higher than that of water processes because 87% of net soil redistribution was linked to tillage. This confirmed the importance of including landscape structure and working at the landscape scale rather than at the plot scale to better estimate soil redistribution in agricultural areas.

Present address for the author M. Lacoste : INRA, UR 0272 Science du sol, Centre de recherched'Orléans, CS 40001 Ardon, 45075 Orléans Cedex 2; Soil organic carbon (SOC) is a key element of agroecosystems functioning and has a crucial impact on global carbon storage. At the landscape scale, SOC spatial variability is strongly affected by natural and anthropogenic processes and linear anthropogenic elements (such hedges or ditches). This study aims at mapping SOC stocks distribution in the A-horizons for a depth up to 105 cm, at a high spatial resolution, for an area of 10 km2 in a heterogeneous agricultural landscape (North-Western France). We used a data mining tool, Cubist, to build rule-based predictive models and predict SOC content and soil bulk density (BD) from a calibration dataset at 8 standard layers (0 to 7.5 cm, 7.5 to 15 cm, 15 to 30 cm, 30 to 45 cm, 45 to 60 cm, 60 to 75 cm, 75 to 90 cm and 90 to 105 cm). For the models calibration, 70 sampling locations were selected within the whole study area using the conditioned Latin hypercube sampling method. Two independent validation datasets were used to assess the performance of the predictive models: (i) at landscape scale, 49 sampling locations were selected using stratified random sampling based on a 300-m square grid; (ii) at hedge vicinity, 112 sampling locations were selected along transects perpendicular to 14 purposively chosen hedges. Undisturbed samples were collected at fixed depths and analysed for BD and SOC content at each sampling location and continuous soil profiles were reconstructed using equal-area splines. Predictive environmental data consisted in attributes derived from a light detection and ranging digital elevation model (LiDAR DEM), geological variables, land use data and a predictive map of A-horizon thickness. Considering the two validation datasets (at landscape scale and hedge vicinity), root mean square errors (RMSE) of the predictions, computed for all the standard soil layers (up to a depth of 105 cm), were respectively 7.74 and 5.02 g kg− 1 for SOC content, and 0.15 and 0.21 g cm− 3 for BD. Best predictions were obtained for layers between 15 and 60 cm of depth. The SOC stocks were calculated over a depth of 105 cm by combining the prediction of SOC content and BD. The final maps show that the carbon stocks in the soil below 30 cm accounted for 33% of the total SOC stocks on average. The whole method produced consistent results between the two predicted soil properties. The final SOC stocks maps provide continuous data along soil profile up to 105 cm, which may be critical information for supporting carbon policy and management decisions.

Quantitative prediction of soil carbon (C) in the landscape can be achieved by empirical or mechanistic models, or a combination of both. The empirical approach called digital soil mapping, usually involves: collection of a database of soil carbon observations over an area of interest; compilation of relevant covariates for the area; calibration or training of a spatial prediction function based on the observed dataset; interpolation and/or- extrapolation of the prediction function over the whole area; and finally validation using existing or independent datasets. The resulting digital maps of C can be used in landscape mechanistic models simulating soil organic C evolution laterally and vertically (within the profile). Here we demonstrate the two approaches in predicting C stock evolution in a landscape in Northwest of France. We introduce the pedogeomorphometry approach which can combine the two approaches to map soil carbon dynamics at the landscape scale.

International audience; We propose to estimate the Organic Carbon (OC) stocks for the O soil layers, at the national scale. First, the determinants of the OC content variability in soil O-horizons were examined, then the OC content was mapped at the national scale and this map was used to estimate the OC stocks for French forest soils. Three soil datasets were used. Two were national datasets provided by national soil survey programs for soils knowledge and inventory and soil quality survey (IGCS and RMQS). One was provided by the Level I of the European program BioSoil. Various covariates were used to identify the determinants of OC stocks in O-horizons: the soil descriptions provided by the pre-mentioned datasets, with collection of biological, chemical and physical properties of O-horizons and soils, plus a description of surrounding landscapes and climate properties. For OC content mapping in O-horizons, spatially exhaustive covariates were used, such as a digital elevation model (SRTM DEM, 90-m grid) and its derived attributes, the French soils map (1:1,000,000), climatic and land use data. Predictive models of OC stocks in the O-horizons were created using Generalized Boosted Regression Models. Estimation of uncertainties has been afterwards modeled using fuzzy k-means clustering. The final OC stocks in O soil layers were estimated to 15 TgC.

International audience; This paper presents and discusses a review on the availability of the different of sources of information to build the GlobalSoilMap products at the scale of France. We investigated the content of the national soil database which gathered the data produced in the framework of the GIS Sol programs (e. g. IGCS, RMQS, BioSoil and BDAT). Two types of spatial information have driven our work: point data and polygon data. Firstly, we present for a set of GlobalSoilMap soil properties the number of soil profiles that fulfill the specifications for 6 depth intervals using density maps. Secondly, maps of the state of progress of the soil mapping and monitoring programs are discussed. Our results show that even if GlobalSoilMap products are potentially feasible with a few statistical challenges, the spatial and semantic quality will be indubitably heterogeneous.

National audience

Les outils de la Cartographie Numérique des Sols (CNS) permettent de prédire des propriétés pédologiques sur de vastes étendues à partir d’informations limitées sur les sols et de variables exhaustives traduisant les facteurs de formation des sols. L’objectif de cette étude était d’établir, à l’échelle de la région Bretagne, des cartes de prédiction de l’hydromorphie et des zones humides par apprentissage automatique, et de les valider avec des données indépendantes. La première étape a consisté à établir un modèle basé sur les relations sol-environnement physique pour prédire l’hydromorphie selon 4 classes. La méthode appliquée est une classification supervisée par arbre stochastique optimisée (algorithme AdaBoost-SAMME implémenté dans la boîte à outil ‘adabag’ du logiciel R). Le modèle a été calibré à partir de 1652 points où l’hydromorphie était connue et de 14 variables environnementales exhaustives sur la région, puis extrapolé à l’ensemble de la zone d’étude (27 360 km²). Dans une seconde étape, l’extension des zones humides potentielles a été dérivée de celle de l’hydromorphie par reclassement des valeurs prédites selon les critères pédologiques de l’arrêté du 1er octobre 2009 relatif à l’identification et à la délimitation des zones humides. Les prédictions ont été validées à partir d’informations pédologiques (données ponctuelles, cartes pédologiques précises numérisées et référentiel régional pédologique de Bretagne à 1/250 000) et d’inventaires de zones humides considérés comme fiables disponibles dans le Finistère. Les variables qui contribuent le plus au modèle de prédiction de l’hydromorphie sont l’occupation du sol, le matériau parental et la pluviométrie, suivies de la courbure verticale et de la dénivelée au cours d’eau. La précision interne du modèle est satisfaisante : le taux de prédictions correctes calculé à partir des données de calibration est de 77 %, l’indice de Kappa de 70 % et les 4 classes d’hydromorphie sont correctement prédites. La distribution spatiale des sols hydromorphes et des zones humides est très cohérente avec les connaissances préexistantes bien que l’hydromorphie et les zones humides semblent surestimées. Cette surestimation est confirmée par la validation des résultats du modèle par comparaison aux observations ponctuelles indépendantes (n =3 348) et aux cartes des sols précises : les taux de prédictions correctes sont de 55 et 36 % respectivement, et les sols non hydromorphes sont sous-estimés. Cette étude a permis d’établir une procédure complète de prédiction de l’hydromorphie et de l’extension spatiale des zones humides potentielles à l’échelle d’une région, à une résolution fine et assortie une incertitude. La procédure est répétable et adaptable, ce qui permet d’envisager d’améliorer les performances du modèle, voire de la transposer à d’autres secteurs qui disposeraient de suffisamment de données pédologiques et de variables prédictives exhaustives.

Soils and landscapes evolve simultaneously. Soil evolution is controlled by redistribution and transformation processes influenced by topographic and climatic parameters, with also a major contribution of management strategies. The perennial landscape features have a strong influence on soil spatial distribution (geometry) and soil genesis. Building landscapes which enhance soil resilience to degradation processes and increase soil services appears as a promising way to adapt to forthcoming climatic and land use evolutions. The presentation aims to synthetize major results from a research program nicknamed Landsoil which focused on the evolution of agricultural soils over medium time scales (decades to centuries) in relation to changing conditions of land use and climate. Precise study of the soil 3D organization in three contrasted landscapes (Brittany, Touraine, Languedoc-Roussillon) enabled to link soil redistribution in space to landscape components (field geometry, hedges or ditches network) and their past evolution. A dynamic and high resolution spatial modeling approach was developed coupling erosion processes and soil organic matter evolution and was calibrated over past evolution using dating techniques (Cs137, C14, OSL). The resulting Landsoil model was afterwards applied in a prospective manner under different scenarios of land use and climate change over the 21th century. Indicators of soil vulnerability and soil resilience were defined and tested by the comparison of several prospective scenarios applied on a same landscape and by comparison of the contrasted landscapes

International audience; Soil organic carbon (SOC) is a key element of agroecosystems functioning and has a crucial impact on global carbon storage at world scale. SOC spatial variability and temporal dynamics are strongly affected by natural and anthropogenic processes occurring at the landscape scale, such as soil redistribution in the lateral and vertical dimensions by tillage and erosion processes. This study aims at modeling SOC distribution in A-horizons, at high spatial resolution, for an area of 1 000 ha in a complex agricultural landscape (NW France). The study site is characterized by high short distance heterogeneity due to an important diversity of soils (with varying redoximorphic conditions, depth), soil parent material (Aeolian loam cover, granite, hard and soft schist), topography, land use (annual crops, temporary or permanent grasslands) and hedge density. We used learning methods based on soil point data, characterized by soil description, SOC content and bulk density measurements. 200 points were selected using conditioned Latin hypercube sampling in order to cover the whole range of ancillary variables (elevation, Modified Compound Topographic Index, K emissions and land use). This sampling strategy enables to select a limited number of sampling sites coverering the study site heterogeneity. Additive sampling was designed to investigate SOC distribution near hedges (112 points sampled at fixed distances along 14 transects crossing hedges). Predictive environmental data consisted in the data used in the conditioned Latin hypercube sampling, at which were added topographic attributes derivated from the DEM and geological variables. We will discuss the ability of our model to capture and predict the SOC, considering the general SOC distribution trend at the landscape scale and the finer SOC distribution in landscape, at hedgerow proximity. The SOC 3-D map obtained will be used as soil data input in a soil evolution model, coupling SOC dynamics and soil erosion modeling.

Digital soil mapping can be helpful in providing pedological information over wide areas where existing soil information is limited. The aim of this study was to predict soil properties at a regional scale by parametrizing soil-landscape models using a machine-learning method recently applied to soil science concerns: boosted classification and regression trees. To examine soil properties interdependence, a two-step approach was tested: first soil parent material (PM), including bedrock formations and superficial deposits, was predicted; then, predicted PM was included as a predictive variable to estimate natural soil drainage (SD). Others predictive variables included environmental data representing known soil-forming factors: terrain attributes (elevation, slope, profile and plan curvatures, sub-watershed hillslope length, hydrological distance from the nearest stream, aspect, relative elevation above the nearest stream and a Compound Index initially proposed by Beven and Kirkby (1979) and modified by Merot et al. (1995)), geological data, airborne gamma-ray spectrometry (K:Th ratio, deviation from mean K emissions of the related lithological unit) and landscape data (derived from remotely sensed data). The study area is located in Brittany (northwestern France) and covers 4645 km². The training dataset was constructed from existing detailed soil maps (scale 1:25,000) available for 11% of the study area. An additional set of 1148 punctual soil observations spread over the study area represented an independent validation dataset. Based on 20,000 randomly selected pixels from the training area, PM and SD were predicted with overall accuracies of 73 and 70% respectively. While calculated on punctual observations, correct agreement between prediction and observation decreased to 49% for PM and 52% for SD. Predicted PM was the most influential variable for SD prediction, illustrating the relevance of the two-step approach tested. Boosted classification tree appeared to be a particularly adequate and robust procedure for predicting soil properties. Probability of occurrence of the predicted PM was demonstrated to be a relevant indication of prediction quality, allowing distinction between well-predicted and poorly-predicted situations.

absent

A machine learning system (MART) has been used to predict soil parent material (SPM) at the regional scale with a 50-m resolution. The use of point-specific soil observations as training data was tested as a replacement for the soil maps introduced in previous studies, with the aim of generating a more even distribution of training data over the study area and reducing information uncertainty. The 27,020-km2 study area (Brittany, northwestern France) contains mainly metamorphic, igneous and sedimentary substrates. However, superficial deposits (aeolian loam, colluvial and alluvial deposits) very often represent the actual SPM and are typically under-represented in existing geological maps. In order to calibrate the predictive model, a total of 4920 point soil descriptions were used as training data along with 17 environmental predictors (terrain attributes derived from a 50-m DEM, as well as emissions of K, Th and U obtained by means of airborne gamma-ray spectrometry, geological variables at the 1:250,000 scale and land use maps obtained by remote sensing). Model predictions were then compared: i) during SPM model creation to point data not used in model calibration (internal validation), ii) to the entire point dataset (point validation), and iii) to existing detailed soil maps (external validation). The internal, point and external validation accuracy rates were 56%, 81% and 54%, respectively. Aeolian loam was one of the three most closely predicted substrates. Poor prediction results were associated with uncommon materials and areas with high geological complexity, i.e. areas where existing maps used for external validation were also imprecise. The resultant predictive map turned out to be more accurate than existing geological maps and moreover indicated surface deposits whose spatial coverage is consistent with actual knowledge of the area. This method proves quite useful in predicting SPM within areas where conventional mapping techniques might be too costly or lengthy or where soil maps are insufficient for use as training data. In addition, this method allows producing repeatable and interpretable results, whose accuracy can be assessed objectively.

Digital soil mapping as a support to wetland delineation at wetland scale

A machine learning system (MART) was used to predict soil parent material at regional scale. We tested the use of punctual soil observations as training data, instead of soil maps used in previous studies, which enables a more even distribution of the training data over the study area (Brittany, NW France). This region shows mainly metamorphic, igneous and sedimentary substrates. However, superficial deposits (Aeolian loam, colluvium and alluvial material), represent very often the actual soil parent material and are very poorly delimitated by existing geological maps. To calibrate a predictive model, MART requires (i) training data, which correspond to the variable to predict, and (ii) environmental predictors. Model predictions were compared (i) to punctual data not used to calibrate the model (crossvalidation), (ii) to the whole punctual dataset, and (iii) to existing soil maps (external validation). The three kinds of validation showed an accuracy of parent material prediction of 56%, 80% and 54%, respectively. Aeolian loam was among the three best predicted substrates. Poor prediction was associated with uncommon materials and areas with high geological complexity, where existing maps used for external validation were particularly imprecise. Parent material prediction can be use to predict other soil properties, like soil waterlogging

La compréhension du déterminisme des émissions de N2O, troisième gaz à effet de serre d’origine anthropique, constitue un enjeu environnemental majeur. Ce manuscrit présente les travaux de recherche réalisés dans le cadre d’un stage de Master 2 à l’INRA (Orléans) sur l’étude du lien entre dynamique des émissions de N2O et fonctionnement hydrique d’un sol cultivé fertilisé. Cette étude a été réalisée sur une maquette expérimentale de sol de 10 m², mise en pente sous un simulateur de pluie. Les émissions de N2O ont été mesurées pendant 70 jours en 16 points de la maquette, à l’aide d’une chambre mobile et d’un spectromètre laser SPIRIT. Les variables du sol (substrats azotés, pH, masse volumique, teneur en eau) ont été mesurées en amont, milieu et aval du dispositif au cours de l’expérimentation. Sur toutes les zones, un premier pic d’émission a été observé en réponse à une pluie de 24 mm consécutive à un apport en substrat azoté – deux conditions favorisant la dénitrification -, et un second pic, plus faible, a été observé après une pluie de 8 mm. La dynamique des processus hydriques (écoulements latéraux, accumulation d’eau en bas de pente), explique des émissions plus fortes en aval., Understanding the dynamics of N2O emissions, the third anthropogenic greenhouse gas, is a major environmental issue. This manuscript presents the research work carried out as part of a Master 2 internship at INRA (Orléans). It deals with the analysis of the relationships between N2O emissions dynamics, and hydric functioning of a cultivated and fertilised soil. A 10 m² sloped soil model was installed under a rain simulator. N2O emissions were measured for 70 days at 16 points of the model, using a fast-box chamber and a SPIRIT laser spectrometer. Soil characteristics (nitrogen, pH, bulk density, water content) were measured upstream, mid- and downstream during the experiment. On all points, a first N2O peak was observed after a 24 mm rainfall following a nitrogen application - two conditions favoring denitrification -, and a second but lower peak was observed after a 8 mm rainfall. The dynamics of the water processes (lateral flows, accumulation of water at the bottom of the slope), explains stronger emissions at the downstream position.

Projet Scansol; Master