Search

[Departement_IRSTEA]Territoires [TR1_IRSTEA]SYNERGIE [Axe_IRSTEA]TETIS-ATTOS; International audience; Finding suitable models of canopy reflectance in forward simulation mode is a prerequisite for their use in inverse mode to characterize canopy variables of interest, such as leaf area index (LAI) or chlorophyll content. In this study, the accuracy of the three-dimensional reflectance model DART (Discrete Anisotropic Radiative Transfer) was assessed for canopies of different genotypes of Eucalyptus, having distinct biophysical and biochemical characteristics, to improve the knowledge on how these characteristics are influencing the reflectance signal as measured by passive orbital sensors. The first step was to test the model suitability to simulate reflectance images in the visible and near infrared. We parameterized DART model using extensive measurements from Eucalyptus plantations including 16 contrasted genotypes. Forest inventories were conducted and leaf, bark, and forest floor optical properties were measured. Simulation accuracy was evaluated by comparing the mean top of canopy (TOC) bidirectional reflectance of DART with TOC reflectance extracted from a Pleiades very high resolution satellite image. Results showed a good performance of DART with mean reflectance absolute error lower than 2%. Intergenotype reflectance variability was correctly simulated, but the model did not succeed at catching the slight spatial variation for a given genotype, excepted when large gaps appeared due to tree mortality. The second step consisted of sensitivity analysis to explore which biochemical or biophysical characteristics influenced more the canopy reflectance between genotypes. Perspectives for using DART model in inversion mode in these ecosystems were discussed. [ABSTRACT FROM PUBLISHER]

International audience; Many small islands are located in the southwestern Indian Ocean. These islands have their own environmental specificities and very fragmented landscapes. Generic land use products developed from low and medium resolution satellite images are not suitable for studying these small territories. This is why we have developed a land use/land cover product, called Homisland-IO, based on remote sensing processing on high spatial resolution satellite images acquired by SPOT 5 satellite between December 2012 and July 2014. This product has been produced using an object-based classification process. The overall accuracy of the product is 86%. Homisland-IO is freely accessible through a web portal and is thus available for future use.

International audience; Timely and efficient land-cover mapping is of high interest, especially in agricultural landscapes. Classification based on satellite images over the season, while important for cropland monitoring, remains challenging in subtropical agricultural areas due to the high diversity of management systems and seasonal cloud cover variations. This work presents supervised object-based classifications over the year at 2-month time-steps in a heterogeneous region of 12,000 km(2) in the Sao Paulo region of Brazil. Different methods and remote-sensing datasets were tested with the random forest algorithm, including optical and radar data, time series of images, and cloud gap-filling methods. The final selected method demonstrated an overall accuracy of approximately 0.84, which was stable throughout the year, at the more detailed level of classification; confusion mainly occurred among annual crop classes and soil classes. We showed in this study that the use of time series was useful in this context, mainly by including a small number of highly discriminant images. Such important images were eventually distant in time from the prediction date, and they corresponded to a high-quality image with low cloud cover. Consequently, the final classification accuracy was not sensitive to the cloud gap-filling method, and simple median gap-filling or linear interpolations with time were sufficient. Sentinel-1 images did not improve the classification results in this context. For within-season dynamic classes, such as annual crops, which were more difficult to classify, field measurement efforts should be densified and planned during the most discriminant window, which may not occur during the crop vegetation peak.

[Departement_IRSTEA]Territoires [TR1_IRSTEA]SYNERGIE [Axe_IRSTEA]TETIS-ATTOS [TR2_IRSTEA]SYNERGIE [ADD1_IRSTEA]Équiper l'agriculture [ADD2_IRSTEA]Dynamiques spatiales d'anthropisation; International audience; The recently launched Sentinel-1 satellite with a Synthetic Aperture Radar (SAR) sensor onboard offers a powerful tool for irrigation monitoring under various weather conditions, with high spatial and temporal resolution. This research discusses the potential of different metrics calculated from the Sentinel-1 time series for mapping irrigated fields. A methodology for irrigation mapping using SAR data is proposed. The study is performed using VV (vertical-vertical) and VH (vertical-horizontal) polarizations over an agricultural site in Urgell, Catalunya (Spain). With field segmentation information from SIGPAC (the Geographic Information System for Agricultural Parcels), the backscatter intensities are averaged within each field. From the Sentinel-1 time series for each field, the statistics and metrics, including the mean value, the variance of the signal, the correlation length, and the fractal dimension, are analyzed. With the Support Vector Machine (SVM), the classification of irrigated crops, irrigated trees, and non-irrigated fields is performed with the metrics vector. The results derived from the SVM are validated with ground truthing from SIGPAC over the whole study area, with a good overall accuracy of 81.08%. Random Forest (RF) machine classification is also tested in this study, which gives an accuracy of around 82.2% when setting the tree depth at three. The methodology is based only on SAR data, which makes it applicable to all areas, even with frequent cloud cover, but this method may be less robust when irrigation is less dominated to soil moisture change.

As planícies de inundação amazônicas são áreas de grande importância econômica, social e ecológica. A ocupação desses ambientes complexos e frágeis traz variados impactos. Esse estudo pretende analisar dinâmicas de alteração de uso e cobertura da terra em uma região de planície de inundação, próximo a cidade de Santarém (PA), sendo a bacia hidrográfica do Lago Grande do Curuai o foco da pesquisa. Para isso foram utilizadas geotecnologias associadas à métricas de paisagens, e uma análise a partir de dois recortes geográficos distintos na mesma área de estudo. O primeiro recorte geográfico é equivalente a toda a bacia hidrográfica, tendo foco na análise da cobertura vegetal natural, correspondente as florestas de terra firme e de várzea; o segundo recorte diz respeito a divisão da área de estudo em seis zonas de paisagem, sendo que essa divisão levou em conta os sistemas produtivos e as características de uso da região em sua delimitação. Foram classificadas imagens dos satélites Landsat 5 e 7, dos anos de 1985, 1997 e 2014. As métricas de paisagem permitiram estabelecer padrões da paisagem, facilitando a comparação entre os momentos distintos. No primeiro recorte, as métricas apontam para uma dinâmica moderada no período; embora haja uma diminuição das áreas de floresta de terra firme, a classe em que há maior aumento é a de vegetação secundária/capoeira. Ocorrem aumentos das áreas de pastagem, entretanto as áreas de cultivo de roça mantêm-se estáveis. As áreas de pasto e roça ocorrem de forma dispersa no território e em pequenas parcelas; essa dinâmica apontada pela classificação é um indicativo de que as áreas de cultivo ocorrem de forma rotacionada, com a ocupação de áreas diferentes que vão se recuperando com o passar do tempo. No segundo recorte, dentre as seis zonas da paisagem, duas são concentradoras de grande parte dos processos de substituição da cobertura vegetal natural por atividades antrópicas, sendo as zonas 3 e 4 as que tem mais alteração da paisagem. A análise em recortes geográficos distintos permitiu identificar que a área de estudo possui padrões heterogêneos de ocupação no tempo e no espaço. Floodplains in the Amazon are areas of great economic, social and ecological importance. Occupation in these areas brings various impacts due to its complexity and fragility. This paper aims to analyze dynamics of change in land use and land cover in the floodplain of Lago Grande do Curuai, near Santarém city, in the state of Pará. To that end, geotechnologies associated with landscape metrics and an analysis of two different geographic clippings of the region were used. The first geographic clipping covers the entire hydrographic basin, focusing on the analysis of the natural vegetation cover (dryland and lowlands); the second geographic clipping shows the area divided in six landscape zones, whose criteria are the productive systems and the characteristics of use of the region. We classified images from satellites Landsat 5 and 7 from the years 1985, 1997 and 2014. Landscape metrics established patterns that facilitated the comparison between different times. In the first geographic clipping, metrics show a moderate dynamic in the period. Although there is a decrease in dry land forest areas, the class in which there is greater increase is the secondary/"capoeira" vegetation. There are increases in the pasture areas, the farmlands, however, remain stable; pasture and farming areas occur in a dispersed way in the territory and in small allotments; this dynamic revealed by classification indicates that cultivation areas are occupied by rotation, being occupied and abandoned over time. In the second geographic clipping, zones 3 and 4 concentrate most of the processes of replacing the natural vegetation cover by anthropic activities. The analysis in different geographic clippings shows the researched area has heterogeneous patterns of occupation in time and space. Les bassins d’inondation amazoniens sont de grande valeur économique, sociale et écologique. Ces espaces complexes et fragiles sont confrontés à une grande diversité d’usage. Cette recherche analyse la dynamique de changement de l’occupation du sol dans le bassin d’inondation du Lac Grand do Curuai, à proximité de Santarém (État du Pará). Pour cela, des métriques paysagères ont été calculées dans deux zonages préalablement sélectionnés afin d’affiner l’analyse des changements spatio-temporels. D’une part les traitements ont porté sur l’ensemble du bassin versant afin d’appréhender les différents types de végétation naturelle (forêt de varzea et forêt de terre ferme). D’autre part, il s’est agit d’effectuer les calculs dans un zonage paysager qui exprime la diversité des systèmes agricoles associés aux caractéristiques environnementales de la région. Les images de satellites utilisées (Landsat 5 et Landsat 7) permettent de conduire l’analyse sur 3 périodes : 1985, 1997 et 2014. Les métriques paysagères ont aidé à établir un modèle paysager qui a pour fonction la comparaison dans le temps et dans l’espace. Les calculs effectués dans le premier zonage expriment une dynamique d’évolution modérée. Si la classe de « forêt de terre ferme » diminue, celles de « végétation secondaire» et de « pâturage » augmentent en superficie. En revanche, la classe « parcelle de manioc » reste stable. Ces classes dispersées dans la zone d’étude, sont de petite taille et changent d’emplacement d’une période sur l’autre. Cette répartition et ces changements sont le signe d’un usage rotationnel des terres. Le deuxième niveau de zonage montre que ces évolutions rotationnels ont essentiellement lieu dans deux des paysages les plus anthropisés. Les dynamiques des autres paysages augurent des modifications d’usage des milieux. L'analyse basée sur ces deux types de zonages permet de mieux affiner les hétérogénéités spatiales dans le temps et dans l'espace.

The research and improvement of methods to be used for crop monitoring are currently major challenges, especially for radar images due to their speckle noise nature. The European Space Agency’s (ESA) Sentinel-1 constellation provides synthetic aperture radar (SAR) images coverage with a 6-day revisit period at a high spatial resolution of pixel spacing of 20 m. Sentinel-1 data are considerably useful, as they provide valuable information of the vegetation cover. The objective of this work is to study the capabilities of multitemporal radar images for rice height and dry biomass retrievals using Sentinel-1 data. To do this, we train Sentinel-1 data against ground measurements with classical machine learning techniques (Multiple Linear Regression (MLR), Support Vector Regression (SVR) and Random Forest (RF)) to estimate rice height and dry biomass. The study is carried out on a multitemporal Sentinel-1 dataset acquired from May 2017 to September 2017 over the Camargue region, southern France. The ground in-situ measurements were made in the same period to collect rice height and dry biomass over 11 rice fields. The images were processed in order to produce a radar stack in C-band including dual-polarization VV (Vertical receive and Vertical transmit) and VH (Vertical receive and Horizontal transmit) data. We found that non-parametric methods (SVR and RF) had a better performance over the parametric MLR method for rice biophysical parameter retrievals. The accuracy of rice height estimation showed that rice height retrieval was strongly correlated to the in-situ rice height from dual-polarization, in which Random Forest yielded the best performance with correlation coefficient R 2 = 0.92 and the root mean square error (RMSE) 16% (7.9 cm). In addition, we demonstrated that the correlation of Sentinel-1 signal to the biomass was also very high in VH polarization with R 2 = 0.9 and RMSE = 18% (162 g·m − 2 ) (with Random Forest method). Such results indicate that the highly qualified Sentinel-1 radar data could be well exploited for rice biomass and height retrieval and they could be used for operational tasks.

Global wheat production reached 754.8 million tons in 2017, according to the FAO database. While wheat is considered as a staple food for many populations across the globe, mapping wheat could be an effective tool to achieve the SDG2 sustainable development goal&mdash, End Hunger and Secure Food Security. In Lebanon, this crop is supported financially, and sometimes technically, by the Lebanese government. However, there is a lack of statistical databases, at both national and regional scales, as well as critical information much needed in the subsidy and compensation system. In this context, this study proposes an innovative approach, named Simple and Effective Wheat Mapping Approach (SEWMA), to map the winter wheat areas grown in the Bekaa plain, the primary wheat production area in Lebanon, in the years of 2016 and 2017. The proposed methodology is a tree-like approach relying on the Normalized Difference Vegetation Index (NDVI) values of four-month period that coincides with several phenological stages of wheat (i.e., tillering, stem extension, heading, flowering and ripening). The usage of the freely available Sentinel-2 imageries, with a high spatial (10 m) and temporal (5 days) resolutions, was necessary, particularly due to the small sized and overlapped plots encountered in the study area. Concerning the wheat areas, results show that there was a decrease from 11,063 &plusmn, 1309 ha in 2016 to 7605 &plusmn, 1184 in 2017. When SEWMA was applied using 2016 ground truth data, the overall accuracy reached 87.0% on 2017 data, whereas, when implemented using 2017 ground truth data, the overall accuracy was 82.6% on 2016 data. The novelty resides in executing early classification output (up to six weeks before harvest) as well as distinguishing wheat from other winter cereal crops with similar NDVI yearly profiles (i.e., barley and triticale). SEWMA offers a simple, yet effective and budget-saving approach providing early-season classification information, very crucial to decision support systems and the Lebanese government concerning, but not limited to, food production, trade, management and agricultural financial support.

[Departement_IRSTEA]Territoires [TR1_IRSTEA]SYNERGIE [Axe_IRSTEA]TETIS-ATTOS [ADD1_IRSTEA]Dynamiques spatiales d'anthropisation; International audience; The objective of this paper is to evaluate the potential of Sentinel-1 Synthetic Aperture Radar "SAR" data (C-band) for monitoring agricultural frozen soils. First, investigations were conducted from simulated radar signal data using a SAR backscattering model combined with a dielectric mixing model. Then, Sentinel-1 images acquired at a study site near Paris, France were analyzed using temperature data to investigate the potential of the new Sentinel-1 SAR sensor for frozen soil mapping. The results show that the SAR backscattering coefficient decreases when the soil temperature drops below 0 °C. This decrease in signal is the most important for temperatures that ranges between 0 and -5 °C. A difference of at least 2 dB is observed between unfrozen soils and frozen soils. This difference increases under freezing condition when the temperature at the image acquisition date decreases. In addition, results show that the potential of the C-band radar signal for the discrimination of frozen soils slightly decreases when the soil moisture decreases (simulated data were used with soil moisture contents of 20 and 30 vol%). The difference between the backscattering coefficient of unfrozen soil and the backscattering coefficient of frozen soil decreases by approximately 1 dB when the soil moisture decreases from 30 to 20 vol%). Finally, the results show that both VV and VH allow a good detection of frozen soils but the sensitivity of VH is higher by approximately 1.5 dB. In conclusion, this study shows that the difference between a reference image acquired without freezing and an image acquired under freezing conditions is a good tool for detecting frozen soils.

[Departement_IRSTEA]Territoires [TR1_IRSTEA]SYNERGIE [Axe_IRSTEA]TETIS-ATTOS [ADD1_IRSTEA]Équiper l'agriculture; International audience; This study evaluates the accuracy of several recent remote sensing Surface Soil Moisture (SSM) products at sites in southwestern France. The products used are Soil Moisture Active Passive "SMAP" (level 3: 36 km x 36 km, level 3 enhanced: 9 km x 9 km, and Level 2 SMAP/Sentinel-1: 1 km x 1km), Advanced Scatterometer "ASCAT" (level 2 with three spatial resolution 25 km x 25 km, 12.5 km x 12.5 km, and 1 km x 1 km), Soil Moisture and Ocean Salinity "SMOS" (SMOS INRA-CESBIO "SMOS-IC", SMOS Near-Real-Time "SMOS-NRT", SMOS Centre Aval de Traitement des Données SMOS level 3 "SMOS-CATDS", 25 km x 25 km) and Sentinel-1(S1) (25 km x 25 km, 9 km x 9 km, and 1 km x 1 km). The accuracy of SSM products was computed using in situ measurements of SSM observed at a depth of 5 cm. In situ measurements were obtained from the SMOSMANIA ThetaProbe (Time Domaine reflectometry) network (7 stations between 1 January 2016 and 30 June 2017) and additional field campaigns (near Montpellier city in France, between 1 January 2017 and 31 May 2017) in southwestern France. For our study sites, results showed that (i) the accuracy of the Level 2 SMAP/Sentinel-1 was lower than that of SMAP-36 km and SMAP-9 km; (ii) the SMAP-36 km and SMAP-9 km products provide more precise SSM estimates than SMOS products (SMOS-IC, SMOS-NRT, and SMOS-CATDS), mainly due to higher sensitivity of SMOS to RFI (Radio Frequency Interference) noise; and (iii) the accuracy of SMAP-36 km and SMAP-9 km products was similar to that of ASCAT (ASCAT-25 km, ASCAT-12.5 km and ASCAT-1 km) and S1 (S1-25 km, S1-9 km, and S1-1 km) products. The accuracy of SMAP, Sentinel-1 and ASCAT SSM products calculated using the average of statistics obtained on each site is defined by a bias of about -3.2 vol. %, RMSD (Root Mean Square Difference) about 7.6 vol. %, ubRMSD (unbiased Root Mean Square Difference) about 5.6 vol. %, and R coefficient about 0.57. For SMOS products, the station average bias, RMSD, ubRMSD, and R coefficient were about -10.6 vol. %, 12.7 vol. %, 5.9 vol. %, and 0.49, respectively.

[Departement_IRSTEA]Territoires [TR1_IRSTEA]SYNERGIE [ADD1_IRSTEA]Dynamiques spatiales d'anthropisation [ADD2_IRSTEA]Adaptation des territoires au changement global; International audience; The aim of this study is to analyze the sensitivity of airborne Global Navigation Satellite System Reflectometry (GNSS-R) on soil surface and vegetation cover characteristics in agricultural areas. Airborne polarimetric GNSS-R data were acquired in the context of the GLORI'2015 campaign over two study sites in Southwest France in June and July of 2015. Ground measurements of soil surface parameters (moisture content) and vegetation characteristics (leaf area index (LAI), and vegetation height) were recorded for different types of crops (corn, sunflower, wheat, soybean, vegetable) simultaneously with the airborne GNSS-R measurements. Three GNSS-R observables (apparent reflectivity, the reflected signal-to-noise-ratio (SNR), and the polarimetric ratio (PR)) were found to be well correlated with soil moisture and a major vegetation characteristic (LAI). A tau-omega model was used to explain the dependence of the GNSS-R reflectivity on both the soil moisture and vegetation parameters. This article belongs to the Special Issue Environmental Research with Global Navigation Satellite System (GNSS).

International audience; The development and improvement of methods to map agricultural land cover are currently major challenges, especially for radar images. This is due to the speckle noise nature of radar, leading to a less intensive use of radar rather than optical images. The European Space Agency Sentinel-1 constellation, which recently became operational, is a satellite system providing global coverage of Synthetic Aperture Radar (SAR) with a 6-days revisit period at a high spatial resolution of about 20 m. These data are valuable, as they provide spatial information on agricultural crops. The aim of this paper is to provide a better understanding of the capabilities of Sentinel-1 radar images for agricultural land cover mapping through the use of deep learning techniques. The analysis is carried out on multitemporal Sentinel-1 data over an area in Camargue, France. The data set was processed in order to produce an intensity radar data stack from May 2017 to September 2017. We improved this radar time series dataset by exploiting temporal filtering to reduce noise, while retaining as much as possible the fine structures present in the images. We revealed that even with classical machine learning approaches (K nearest neighbors, random forest, and support vector machines), good performance classification could be achieved with F-measure/Accuracy greater than 86% and Kappa coefficient better than 0.82. We found that the results of the two deep recurrent neural network (RNN)-based classifiers clearly outperformed the classical approaches. Finally, our analyses of the Camargue area results show that the same performance was obtained with two different RNN-based classifiers on the Rice class, which is the most dominant crop of this region, with a F-measure metric of 96%. These results thus highlight that in the near future these RNN-based techniques will play an important role in the analysis of remote sensing time series.This article belongs to the Special Issue High Resolution Image Time Series for Novel Agricultural Applications

Based on a standardized typology, this thesis aims to propose a method for mapping the French Overseasmangroves. Several sets of satellite data are used to (i) determine mangrove extent, (ii) quantify canopy closure,(iii) quantify canopy height, and (iv) identify main species.Determination of mangrove extent was possible by using spectral characteristics of mangroves in the shortwave infrared wavelength observed on the Landsat and Sentinel-2 images.Quantification of canopy closure is based on a fully constrained algorithm (FCLSU) of linear spectral mixtureanalysis of pixels of a very high resolution Pléiades images.Quantification of canopy height was performed using photogrammetric analysis of WorldView-3 stereoscopicpanchromatic images (0.3 meters resolution).Finally, discrimination of main mangrove species was permitted through the use of spectral surveys andcalibration of spectral properties of the different species observed on the satellite images., A partir d'une typologie physionomique standardisée, cette thèse entend proposer une méthode decartographie des mangroves de l'Outre-mer français. Plusieurs jeux de données satellites sont ainsi exploitésen vue (i) de déterminer l'extension des surfaces de mangroves, (ii) quantifier la fermeture de canopée, (iii)quantifier la hauteur de la canopée et (iv) identifier les principales espèces de palétuvier.La détermination des surfaces de mangroves se fait via l'exploitation des particularités spectrales despalétuviers dans le moyen-infrarouge visibles sur les images Landsat et Sentinel-2.La quantification de la fermeture de canopée est basée sur un algorithme doublement contraint (FCLSU) dedémélengeage linéaire des pixels d'images très haute résolution Pléiades.La quantification de la hauteur de canopée a été effectuée quant à elle grâce à l'utilisation d'une analysephotogrammétrique d'un couple d'images panchromatiques stéréoscopiques WorldView-3 (0.3mètres de résolution). Enfin, la discrimination des principales espèces demangroves a été permise grâce à l'exploitation de relevés spectraux effectués sur le terrain et calibrationdes propriétés spectrales des différentes espèces visibles sur les images satellites.

International audience; Earthquake is one of the dominant triggering factors of landslides. Given the wide areas covered by mega earthquake-triggered landslides, their inventory requires development of automatic or semi-automatic methods applied to satellite imagery. A detection method is here proposed for this purpose, to fit with simple datasets; SPOT5 panchromatic images of 5 m resolution coupled with a freely and globally available DEM. The method takes advantage of multi-temporal images to detect changes based on radiometric variations after precise coregistration/orthorectification. Removal of false alarms is then undertaken using shape, orientation and radiometric properties of connected pixels defining objects. 80% of the landslides and 93% of the landslide area are detected indicating small omission errors but 50% of false alarms remain. They are removed using expert based analysis of the inventory. The method is applied to realize the first comprehensive inventory of landslides triggered by the Pisco earthquake (Peru, 15/08/2007, Mw 8.0) over an area of 27,000 km2. 866 landslides larger than 100 m2 are detected covering a total area of 1.29 km2. The area/number distribution follows a power-law with an exponent of 1.63, showing a very particular regime of triggering in this arid environment compared to other areas in the world. This specific triggering can be explained by the little soil cover in the coastal and forearc regions of Peru. Analysis of this database finally shows a major control of the topography (both orientation and inclination) on the repartition of the Pisco-triggered landslides.

International audience; Présentation d'une librairie open source pour traiter des images satellitaires

International audience; Finding suitable models of canopy reflectance in forward simulation mode is a prerequisite for their use in inverse mode to characterize canopy variables of interest, such as leaf area index (LAI) or chlorophyll content. In this study, the accuracy of the three-dimensional reflectance model DART (Discrete Anisotropic Radiative Transfer) was assessed for canopies of different genotypes of Eucalyptus, having distinct biophysical and biochemical characteristics, to improve the knowledge on how these characteristics are influencing the reflectance signal as measured by passive orbital sensors. The first step was to test the model suitability to simulate reflectance images in the visible and near infrared. We parameterized DART model using extensive measurements from Eucalyptus plantations including 16 contrasted genotypes. Forest inventories were conducted and leaf, bark, and forest floor optical properties were measured. Simulation accuracy was evaluated by comparing the mean top of canopy (TOC) bidirectional reflectance of DART with TOC reflectance extracted from a Pleiades very high resolution satellite image. Results showed a good performance of DART with mean reflectance absolute error lower than 2%. Intergenotype reflectance variability was correctly simulated, but the model did not succeed at catching the slight spatial variation for a given genotype, excepted when large gaps appeared due to tree mortality. The second step consisted of sensitivity analysis to explore which biochemical or biophysical characteristics influenced more the canopy reflectance between genotypes. Perspectives for using DART model in inversion mode in these ecosystems were discussed. [ABSTRACT FROM PUBLISHER]

This paper presents an overview of a decade (2002-2012) of studies performed on Soil-Plant-Atmosphere processes, in semi-arid areas (Haouz plain, Morocco). In this period, fifteen in-situ experiments on the dominant irrigated crops were performed, controlling the fluxes exchanged between land surface and atmosphere. The results showed that the physically based SVATs (ICARE, SiSPAT and ISBA) provided the best estimates of surface fluxes. For operational purposes, the FAO-56 approach and SAMIR Software (Satellite Monitoring for Irrigation) give a good estimate of evapotranspiration at field and regional scales, respectively. Finally, thermal infrared data were used in conjunction with SVATs to investigate the possibility of estimating the quantity of irrigation water.

Small reservoirs have developed across semi-arid areas as a low cost solution for millions of rural small holders to harvest scarce water resources. Studies have highlighted limited agricultural water use and low water availability on individual reservoirs, but no information exists on the drought patterns of multiple small reservoirs. Their small size and dispersion prevents individualised hydrological monitoring, while hydrological modelling suffers from rainfall variability and heterogeneity across data sparse catchments and reservoirs. A semi-automated original approach exploiting free, archive Landsat satellite images is developed here for long-term monitoring of multiple ungauged small water bodies. Adapted and tested against significant hydrometric time series on three lakes, the method confirms its potential to monitor water availability on the smallest water bodies (1–10 ha) with a mean RMSE of 20,600 m3 (NRMSE = 26%). Uncertainties from the absence of site-specific and updated surface-volume rating curves were here contained through a power relationship adapted over time for silting based on data from 15 surrounding lakes. Applied to 51 small reservoirs and 546 images over 1999–2014, results highlight the ability of this transposable method to shed light on flood dynamics and allow inter annual and inter lake comparisons of water availability. In the Merguellil upper catchment, in Central Tunisia, results reveal the significant droughts affecting over 80% of reservoirs, confirming the need for small reservoirs to maintain a supplementary irrigation objective only.

International audience; Remote sensing tools (RS) can contribute to a better understanding of the diversity of natural and semi-naturals habitats, their spatial distribution, and their conservation status. RS can also provide a generic set of derived indicators to support local to regional habitat monitoring. Here we propose a set of synthetic continuous textural indices computed from high spatial resolution airborne images for the characterization of vegetation structure in very heterogeneous landscape mosaics. These indices are based on Fourier-based textural ordination (FOTO) of very-high-resolution images. We investigate the relationship between textural indices and a set of common landscape metrics derived from vegetation maps, identifying four strata of interest: bare soil, herbs, low ligneous, and high ligneous. We identify two continuous textural indices, the first one being related to vegetation strata fragmentation and the second being related to the dominance of high ligneous. The combination of these two textural indices with the Normalized Difference Vegetation Index (NDVI) provides a synoptic and accurate overview of the spatial organization of the different vegetation strata. The methodological approach presented herein has a generic value in response to national conservation targets in the context of mapping relevant habitat indicators.This article belongs to the Section Remote Sensing in Agriculture and Vegetation.

My research activities are related to the fields of Data Mining, Data Base and Machine Learning. The main goal of my work is the development of new techniques and algorithms to manage and analyze large amounts of heterogeneous data with a major emphasis on data involving spatial and temporal characteristics (i.e. satellite images, environmental data, sensor data, etc...).

[Departement_IRSTEA]Territoires [TR1_IRSTEA]SYNERGIE [Axe_IRSTEA]TETIS-ATTOS; International audience; The objective of this paper is to propose a new semi-empirical radar backscattering model for bare soil surfaces based on the Dubois model. A wide dataset of backscattering coefficients extracted from synthetic aperture radar (SAR) images and in situ soil surface parameter measurements (moisture content and roughness) is used. The retrieval of soil parameters from SAR images remains challenging because the available backscattering models have limited performances. Existing models, physical, semi-empirical, or empirical, do not allow for a reliable estimate of soil surface geophysical parameters for all surface conditions. The proposed model, developed in HH, HV, and VV polarizations, uses a formulation of radar signals based on physical principles that are validated in numerous studies. Never before has a backscattering model been built and validated on such an important dataset as the one proposed in this study. It contains a wide range of incidence angles (18-57) and radar wavelengths (L, C, X), well distributed, geographically, for regions with different climate conditions (humid, semi-arid, and arid sites), and involving many SAR sensors. The results show that the new model shows a very good performance for different radar wavelengths (L, C, X), incidence angles, and polarizations (RMSE of about 2 dB). This model is easy to invert and could provide a way to improve the retrieval of soil parameters.

International audience; Over 4 years, 2001–2004, leatherback-turtle monitoring was conducted on all the potential nesting sites in French Guiana. We estimated minimal leatherback turtle nest numbers of 23,107, 12,229, 13,480, 11,012, respectively. The Awala-Yalimapo Beach, sometimes considered a good estimator of the overall nesting activity for the country, has hosted a significant proportion of the leatherback turtle nests (42%±2%), but this percentage is much lower than formerly described. The relative importance of this nesting site is discussed in light of remote sensing data, suggesting that nest numbers recorded in Awala-Yalimapo may have misrepresented leatherback turtle population trends. Indeed, remote sensing data indicate that the total sandy shoreline available in French Guiana has regularly evolved over the last decades, allowing leatherback turtle nesting attempts out of the scope of monitoring. The importance of a monitoring effort integrating the specific coastal dynamic of the Guianas region is highlighted.

National audience; Large-scale characterization of tropical rainforest is a challenge for their conservation and management. Very high spatial resolution images as provided by the Pléiades satellites offer new opportunities to study the structural organization of heterogeneous rainforests with limited accessibility. In this study, we have evaluated the potential of Pléiades images to map structural parameters of New Caledonian rainforests by analyzing texture of forest canopies. We have applied the Fourier transform textural ordination (FOTO) method to very high spatial resolution images to compute texture indices of canopy grain (i.e. a combination of size distribution and spatial pattern of tree crowns). The results have showed that this method was promising to isolate the rainforest from other types of vegetation, and to highlight their structural diversity on a large scale. Finally, this case study showed that the use of Pléiades images is promising to predict the structure of rainforests.; Cartographier et classifier les forêts denses humides (FDH) selon une typologie structurale objective est un enjeu majeur pour leur conservation et leur gestion. Les principales contraintes dans l'étude des FDH sont dues à leur forte hétérogénéité et à leur faible accessibilité. Les images satellitaires Pléiades offrent de nouvelles opportunités pour l'étude, à large échelle, de l'organisation structurale de ces forêts. Dans cet article nous avons évalué à travers une étude de cas en Nouvelle-Calédonie leur potentiel pour construire une typologie des FDH à partir d'une analyse de texture de la canopée. La méthode FOTO (FOurier-based Textural Ordination) a été appliquée à des images à très haute résolution spatiale pour produire des indices de texture du grain de la canopée, qui associent la distribution des tailles des couronnes des arbres et leur répartition spatiale. Les résultats ont montré que cette méthode permettait, d'une part d'isoler la FDH des autres formations végétales, et d'autre part de mettre en évidence leur diversité structurale à large échelle. Enfin, cette étude de cas a montré que l'utilisation d'images Pléiades est prometteuse pour prédire la structure des forêts denses humides. Mots-clés : Structure de la canopée, analyse de texture, forêt dense humide, Nouvelle-Calédonie, télédétection, image à très haute résolution Abstract Large-scale characterization of tropical rainforest is a challenge for their conservation and management. Very high spatial resolution images as provided by the Pléiades satellites offer new opportunities to study the structural organization of heterogeneous rainforests with limited accessibility. In this study, we have evaluated the potential of Pléiades images to map structural parameters of New Caledonian rainforests by analyzing texture of forest canopies. We have applied the Fourier transform textural ordination (FOTO) method to very high spatial resolution images to compute texture indices of canopy grain (i.e. a combination of size distribution and spatial pattern of tree crowns). The results have showed that this method was promising to isolate the rainforest from other types of vegetation, and to highlight their structural diversity on a large scale. Finally, this case study showed that the use of Pléiades images is promising to predict the structure of rainforests.

International audience; Connaître la distribution des espèces vectrices de zoonoses est fondamentale pour comprendre les dynamiques épidémiologiques et caractériser les zones à risque. Cette distribution peut être appréhendée à différentes échelles : régionalement, l'aire de répartition de ces espèces définit les limites d'occurrence de la transmission d'agents pathogènes et, localement, leurs niches déterminent l'aléa. Cette dernière échelle requiert une connaissance très précise des milieux qui peut être acquise par l'analyse d'images satellite à très haute résolution spatiale. Dans le cadre d'un projet de recherche sur la leptospirose à La Réunion (LeptOI), cette étude propose de mesurer le potentiel d'utilisation d'une image Pléiades (produit rééchantillonné à 50 cm) pour caractériser la distribution locale de différents petits mammifères terrestres autour de l'Etang de Saint-Paul. Trois campagnes de piégeage ont permis la capture de 134 animaux appartenant à quatre des cinq espèces présentes à La Réunion (Rattus rattus, R. norvegicus, Suncus murinus et Tenrec ecaudatus). Une analyse orientée-objet de l'image Pléiades a permis de discriminer dix classes d'occupation des sols avec une précision totale de 83,6%. Des indices paysagers ont été calculés à partir de cette classification autour de chacun des pièges (distances les plus courtes aux classes d'occupation, densité de contours et pourcentages de surface des classes au sein de zones tampon de 50 et de 100 mètres). Trois zones (« humide », « anthropisée » et « sèche ») sont clairement discriminées par ces variables, qui pourront être utilisées comme déterminants écologiques de différents profils épidémiologiques : la zone humide étant propice à la leptospirose et la zone anthropisée étant un lieu de cooccurrence des quatre espèces et ainsi potentiellement une zone d'échanges de pathogènes. Abstract Knowing the distribution of vectors of zoonoses is fundamental to understand epidemiological dynamics and characterize risk areas. This distribution can be assessed at different scales: regionally, the range of these species defines the limits of occurrence of pathogen transmission and, locally, niches determine the hazard. This latter scale requires a very precise knowledge of the areas that can be gained through the analysis of very high spatial resolution satellite images. As part of a research project on leptospirosis in La Réunion (LeptOI), this study proposes to measure the potential of a Pleiades image (product resampled to to 50 cm) to characterize the local distribution of various small terrestrial mammals in Saint-Paul swamp area. Three trapping campaigns have led to the capture of 134 animals belonging to four of the five species occurring in Reunion Island (Rattus rattus, R. norvegicus, Suncus murinus and Tenrec ecaudatus). An object-based image analysis of the Pleiades image allowed discriminating ten land use classes with a total accuracy of 83.6%. Landscape indices were calculated from the classification around each trap (shortest distances to each land use class, edge densities and proportion of each patch within buffers defined by a 50 meters or a 100 meters radius). Three areas ("wet", "urbanized" and "dry") are clearly discriminated by these variables, which can be used as ecological determinants of different epidemiological profiles. The wetland area is known as suitable to leptospirosis whereas the urbanized area is a place of co-occurrence of the four species and thus potentially favorable to the exchange of pathogens.

Due to the El Nino phenomenon, the 1997-1998 dry season in Roraima (Brazil, Amazonia) was particularly pronounced. Consequently, vegetation fires spread widely and were monitored by many satellites in real time. Satellite images are currently being used to monitor vegetation fires either globally for climate studies or more regionally for impact assessment. After reviewing different satellite data used for impact assessment, this paper focuses on the contribution of SPOT-4's imagery provided by high resolution HRVIR and coarse resolution VEGETATION sensors. These sensors are described with emphasis on those characteristics of potential benefit for forest mapping and fire detection. Early images of Roraima from SPOT-4 are analysed and interpreted to delineate the areas already damaged by fire. VEGETATION images provide a first estimate of damaged areas on a regional scale and an indication of the main ecosystems affected. SPOT HRVIR is used to establish a much more precise classification of various ecosyst...

To handle the integration of intermittent energies to the existing grid,managers require more and more acurate tools to forecast the primary resources. This thesisfocuses on the very short term forecast of the global horizontal irradiance (GHI), maininput for most photovoltaic power plant. We chose to use both ground based images fromhemispherical cameras and satellite images to provide a forecating tool. In the first handwe present a novel appraoch to estimate the GHI using ground based images. On secondhand, we propose several satellite based methods to forecast the GHI up to one hour. Finally,we developp a new method allowing us to merge both data in order to benefit from theirrespective advantages. All the methods were tested against real data acquired on the SIRTAsite, Polytechnique campus., L’intégration des énergies intermittentes sur le réseau électrique existant soulèvedes problèmes de stabilité de la balance consommation/production. Afin de limiter les risques,la loi autorise un taux de pénétration maximum de ces énergies de 30% de la puissanceinstallée. Afin de pallier cette limitation, deux pistes sont envisagées, le stockage d’énergie etla prévision de production. Les travaux menés dans cette thèse s’inscrivent dans la secondecatégorie et se concentrent sur la prévision du rayonnement global au sol, ressource principaledes systèmes de production d’énergie photovoltaïque. Dans l’objectif d’une prévision à trèscourt terme (infra-horaire), la problématique développée concerne la fusion d’informationsprovenant d’une part d’observations satellitaires et de l’autre d’observations in-situ telles quedes images hémisphériques. L’approche suivie se veut progressive et s’articule autour de 4grand axes. Le premier chapitre énonce les définitions et les enjeux principaux liés à l’étudedu GHI en décrivant les différents facteurs physiques ayant des interactions sur sa mesure. Lesecond chapitre permet d’évaluer le potentiel des images hémisphériques pour l’estimation durayonnement global. On y développe une méthode d’estimation du GHI basée sur l’utilisationd’un réseau de neurones artificiels permettant d’effectuer une régression non linéaire entre descaractéristiques choisie de l’image et le rayonnement mesuré sur site par un pyranomètre. Letroisième chapitre concerne l’utilisation des images satellitaires pour la prévision à très courtterme du rayonnement global. L’originalité des méthodes proposées provient de l’utilisationen entrées de cartes de rayonnement dérivées des images satellitaires via une méthode externe.L’utilisation de ces cartes de rayonnement permet la construction de modèles linéairessimples (modèles auto-régressifs) grâce à l’homogénéité des données utilisées. Cependant,leur pertinence pour le calcul de champ de vecteurs a également été prouvé et validé dans unsecond modèle. La comparaison des deux modèles ainsi créés à partir d’imagerie satellitairea permis de caractériser les forces et faiblesses de ces modèles. L’intérêt de l’observationsatellitaire réside dans l’anticipation du déplacement des masses nuageuse. Cependant, unbiais non systématique persiste dans la conversion des valeurs des pixels en rayonnement.Ce biais est inhérent à la résolution spatio-temporelle de ces observations. Étant donné cesconsidérations, le chapitre 4 présente alors une méthode d’intégration des données acquisespar l’imagerie hémisphérique, disposant une meilleure résolution spatio-temporelle, dans lesrésultats de prévision satellitaires précédemment évoqués. On joint alors les caractéristiquesretenues au chapitre 2 dans un réseau de neurone avec la prévision pour produire uneprévision dont le biais est largement réduit. L’utilisation de caractéristiques dérivées del’imagerie hémisphérique à la place de simple mesures du pyranomètre est justifiée par l’effetde persistance non souhaité apportées par ces dernières. Ainsi, après avoir étudié chaquesource d’information séparément, on a pu développer un modèle leur permettant de secompléter.

Le modèle de potentiel a été et reste un modèle d'interaction spatiale utilisé pour diverses problématiques en sciences humaines, cependant l'utilisation qu'en ont fait Donnay (1997,1995,1994) et Binard (1995) en introduisant des résultats de traitement d'images comme support d'application a ouvert la voie à des applications novatrice par exemple, pour la détermination de la limite urbaine ou des hinterlands locaux. Les articulations possibles entre application du modèle de potentiel en imagerie et utilisation de plans de Système d'Information Géographique ont permis l'évaluation temporelle des tendances de développement urbain (Weber,1998). Reprenant cette idée, l'étude proposée tente d'identifier les formes de développement urbain de la Communauté urbaine de Strasbourg (CUS) en tenant compte de l'occupation du sol, des caractéristiques des réseaux de communication, des réglementations urbaines et des contraintes environnementales qui pèsent sur la zone d'étude. L'état initial de l'occupation du sol, obtenu par traitements statistiques, est utilisé comme donnée d'entrée du modèle de potentiel afin d'obtenir des surfaces de potentiel associées à des caractéristiques spatiales spécifiques soit : l'extension de la forme urbaine, la préservation des zones naturelles ou d'agricultures, ou encore les réglementations. Les résultats sont ensuite combinés et classés. Cette application a été menée pour confronter la méthode au développement réel de la CUS déterminé par une étude diachronique par comparaison d'images satellites (SPOT1986- SPOT1998). Afin de vérifier l'intérêt et la justesse de la méthode les résultats satellites ont été opposés à ceux issus de la classification des surfaces de potentiel. Les zones de développement identifiées en fonction du modèle de potentiel ont été confirmées par les résultats de l'analyse temporelle faite sur les images. Une différenciation de zones en fonction des résultats du modèle et de l'intégration du réseau routier a pu être établie. L'intérêt de cette approche réside dans la possibilité de fournir une information spatiale, non accessible à partir de données couramment utilisées en aménagement et qui identifie des zones propices ou non au développement en fonction de choix ou de contraintes énoncées. Interaction models and especially potential model have often been used for demographic or social purposes, in order to study the relationships dimension between population types and locations or between users and goods etc. Defining the interactions between space characteristics, Donnay (1997, 1995, 1994) and Binard (1995) have promoted the application of such model on remote sensing products in order to obtain urban delimitation. Another use of potential model also dealing with remotely sensed product has been proposed by Weber et al. (1997 ; Weber, 1998) by the integration of potential model principles and GIS capacities to evaluate development trends of an urban settlement over years. In this case GIS is used to diversify the importance of locations taken into account over the studied area. This could be defined by spatial information (as initial input) and other characteristics attached : urban regulation, environment protection laws, urban development trends etc. This allows multiplying the current visions of urban area development. Referring to natural constrains or anthropic decisions such approach offers the possibility to integrate different types of interactions between various locations. Statistical methods are applied to get land-cover classification results of the images used, then potential model is processed in order to obtain potential surfaces associated with specific spatial characteristics. The resulting layers might be combined to get composed information on the area or ordered results. The application presented has been conducted on the Strasbourg region in order to assess the method according the city development reality. A time analysis has been set up linked to spatial changes evaluation. Dynamic processes have been identified through sequential estates (1986 – 1998). The potential model application is used to get a more predictive idea of the spatial changes reality, allocating diverse weights to landscape locations. Two satellite images have been used : the first one to get a first idea of the land-cover surface on Strasbourg and the surroundings (SPOT XS 1986), and the second one has been processed (1) to perform time series analysis and (2) to assess the accuracy of the potential model application (SPOT XS 1998). Ancillary data are gathered in order to enrich our knowledge on the studied area through environmental, planning or demographic points of view. The expected results are the identification of development trends from potential model application on remotely sensed results. Point to be stressed is the determination of the location weights according to spatial characteristics. Methods considered to deal with the combination of weights are close to multicriteria analysis.

International audience; The constant migration of mudbanks resulting from the sedimentary discharge of the Amazon River has an enormous impact on the French Guianese economy. Consequently, there is great local interest in finding tools that can be used in mapping and monitoring the mudbank migration with an acceptable level of reliability and a high frequency of measurement. This involves the ability to detect submarine mudbanks that interfere with access to port channels and impede navigation. Radar and optical satellite images of French Guiana acquired between 1997 and 2001 have been analysed to assess their utility in detecting mudbanks and monitoring coastline change. In the present study, the information derived from radar and optical satellite imagery was analysed with respect to tidal height. The influence of the radar incidence angle was considered, as well as the advantages of having multidate sequences when mapping coastal zones undergoing significant change. A comparative study between radar (European Remote Sensing Satellite (ERS) and RADARSAT) and optical (advanced spaceborne thermal emission and reflection radiometer (ASTER)) images was also conducted. The results show that low-angle radar is more suitable for detecting mudbanks, whereas high-angle radar is more appropriate for monitoring coastline change. Emerged mudbanks and mudbanks under a shallow layer of water (a few tens of centimetres at the most) are easily detectable in radar images. For high tidal levels, the optical images provide more information on the mudbanks than the radar images. Conversely, for low tidal levels, the information on the mudbanks is more detailed with radar imagery.; Le déplacement continu des bancs de vase en provenance de l'Amazone a un impact très important sur l'économie guyanaise. La recherche d'outils permettant la cartographie des bancs et le suivi de leur migration dans des conditions acceptables de fiabilité et de rapidité présente donc un grand intérêt local. Ceci se traduit par la détection en mer des bancs de vase qui viennent perturber l'accès aux chenaux (ports) ce qui constitue un gêne pour la navigation. Des images satellitaires radar et optique, acquises sur la Guyane française entre 1997 et 2001, ont été analysées pour examiner leurs apports à la détection des bancs de vase, et au suivi de l'évolution du trait de côte. L'information issue de l'imagerie satellitaire radar et optique a été étudiée en fonction de la hauteur d'eau. Nous avons analysé l'influence de l'angle d'incidence radar et l'intérêt des séquences multi-dates pour cartographier les zones côtières soumises à des importantes modifications. Une étude comparative entre imagerie radar (ERS et RADARSAT) et optique (ASTER) a également été effectuée ainsi que la relation entre bathymétrie et information satellitaire. Les résultats montrent qu'une incidence radar faible est plus adaptée à la détection des bancs de vase tandis qu'une incidence radar forte est plus appropriée pour le suivi du trait de côte. Les bancs de vase émergés ou immergés sous une faible couche d'eau (quelques dizaines de centimètres au maximum) sont facilement détectables sur les images radar. Pour des hauteurs d'eau élevées, l'imagerie optique permet une meilleure caractérisation des bancs de vase comparativement à l'imagerie radar. Dans le cas d'une faible hauteur d'eau, les résultats montrent que les informations sur les bancs de vase sont plus détaillées avec l'imagerie radar.

International audience; Features on the seabed can be mapped from remote sensing multi/hyperspectral imagery, provided that their effects on the measured reflectance spectrum can be made independent of those produced by the atmosphere and water column. The non-linear effect of water column light attenuation can then be corrected to obtain the absolute reflectance of the seabed. Light attenuation by the water column and bathymetry are both determined from the satellite image. The water column attenuation is then removed in order to apply an automated supervised classification, whatever the depth is. We have compared the results obtained with and without the correction of water column attenuation, for two different statistical measures: Euclidean (ED) and spectral angle mapper (SAM) distances. We have applied this methodology to MERIS images acquired on the lagoon of New Caledonia. The best overall accuracy (79%), as compared to in situ data, is obtained with the corrected image and the SAM distance.

La description des structures végétales fournit des informations essentielles à une meilleure compréhension de la dynamique des forêts. Que ce soit à l'échelle de l'arbre, des peuplements ou des massifs forestiers, notre capacité à mesurer et à suivre sur le terrain l'évolution des paramètres structuraux et botaniques qui décrivent cette dynamique reste limitée vu la diversité des associations végétales, la complexité des couverts forestiers, l'étendue des milieux et parfois les difficultés d'accès. Expliquer la croissance des arbres et le développement des forêts nécessiterait une approche intégrée permettant d'appréhender à toutes les échelles spatiales et temporelles d'observation, des processus endogènes et exogènes à la plante et aux peuplements d'arbres. Dans cette problématique, le couplage, sur la base de mesures de terrain pertinentes, de l'observation spatiale dans plusieurs domaines électromagnétiques avec des modèles écologiques et environnementaux me parait un maillon fondamental de la chaîne d'analyse à tenter de mettre en place pour étudier les forêts. C'est le cadre du projet de recherche que je développe au sein de l'UMR AMAP, avec un focus sur les forêts tropicales depuis mon recrutement à l'IRD. Je présenterai la chaîne de traitement et d'analyse que j'ai mise en place et qui part de mesures de terrain de plus en plus fines pour aller jusqu'à l'utilisation d'une modélisation physique permettant de simuler des images optiques de forêt à très haute résolution spatiale. J'illustrerai en donnant des résultats obtenus dans l'évaluation du potentiel des mesures de télédétection, aussi bien radar qu'optique pour, par exemple, suivre le cycle végétatif en forêt tempérée, caractériser une variabilité de structures forestières tropicales, mettre en évidence des processus de colonisation et estimer des biomasses épigées en mangroves. C'est sur ce dernier écosystème que mon travail s'est focalisé peu à peu. Que ce soit sur le terrain ou au travers l'analyse des images de télédétection, l'étude des mangroves ramène constamment à l'hypothèse de Thompson (1917) dans son livre intitulé 'On growth and form": Physical forces, not heredity, may govern biological form. C'est avec une vision à long terme que je tente de développer un chantier interdisciplinaire qui repose sur des collaborations pérennes développées au sein et à l'extérieur de l'UMR AMAP (France & Guyane Française, Allemagne, Brésil, Royaume Uni, Indonésie). Dans ce sens, j'expliquerai comment la description de la structure des forêts de palétuviers peut servir de prisme au travers duquel on peut aborder la simulation de la dynamique des côtes à mangroves. Enfin, je présenterai l'intérêt et, j'espère, les prémices d'un réseau international de stations d'observation des mangroves dans lequel protocoles de mesures, méthodes, modèles seraient partagés.

An estimation of bathymetry in the optically complex lagoon of New Caledonia was performed from multispectral satellite images. Bathymetric estimation was obtained with Medium Resolution Imaging Spectrometer (MERIS) images acquired at different dates. The method is based on the rotation of a pair of spectral bands (red and green bands). One of the resulting images is only depth-dependent when the other is reflectance-dependent. A comparison was made with a compilation of bathymetric data from Institut de Recherche pour le Developpement (IRD) Noumea measurements. This study shows the influence of the variation of water composition, and then of the diffuse attenuation, on bathymetric results. The variation of chlorophyll concentration, correlated with variation of diffuse attenuation, explains the error of the estimation map. The results show that this method remains valid under a low (

Retrievals of inherent optical properties (IOPs) and Chlorophyll-a concentration (Chla) was investigated for AVNIR-2 images with 30-m spatial resolution and four bands in the southwest tropical lagoon of New Caledonia. We corrected the atmospheric and sea-surface reflectance iteratively through the retrieval of IOPs. After an additional correction of sea floor reflectance, the estimated IOPs and Chla agreed well with the in situ measurements even in the lagoon areas. This study provides a method to allowing a local optimal estimation of IOPs and Chla with a high resolution sensor by preparation of the candidate spectra for the target areas.

International audience; In France, in the peri-urban context, urban sprawl dynamics are particularly strong with huge population growth as well as a land crisis. The increase and spreading of built-up areas from the city centre towards the periphery takes place to the detriment of natural and agricultural spaces. The conversion of land with agricultural potential is all the more worrying as it is usually irreversible. The French Ministry of Agriculture therefore needs reliable and repeatable spatial-temporal methods to locate and quantify loss of land at both local and national scales. The main objective of this study was to design a repeatable method to monitor land conversion characterized by artificial sprawl: (i) We used an object-based image analysis to extract artificial areas from satellite images; (ii) We built an artificial patch that consists of aggregating all the peripheral areas that characterize artificial areas. The "artificialized" patch concept is an innovative extension of the urban patch concept, but differs in the nature of its components and in the continuity distance applied; (iii) The diachronic analysis of artificial patch maps enables characterization of artificial sprawl. The method was applied at the scale of four departments (similar to provinces) along the coast of Languedoc-Roussillon, in the South of France, based on two satellite datasets, one acquired in 1996-1997 (Indian Remote Sensing) and the other in 2009 (RapidEye). In the four departments, we measured an increase in artificial areas of from 113,000 ha in 1997 to 133,000 ha in 2009, i. e., an 18% increase in 12 years. The package comes in the form of a 1/15,000 valid cartography, usable at the scale of a commune (the smallest territorial division used for administrative purposes in France) that can be adapted to departmental and regional scales. The method is reproducible in homogenous spatial-temporal terms, so that it could be used periodically to assess changes in land conversion rates in France as a whole.

The European Space Agency (ESA) Climate Change Initiative (CCI) is part of the European contribution to the Global Climate Observing System (GCOS) program. Fire disturbance is one of the Essential Climate Variables (ECV) included in the ESA CCI program. It focus on mapping burned area (BA) using European sensors (ATSR, VEGETATION and MERIS data), and in comparing the performance of the results with other existing datasets. The project aims at developing and validating algorithms to produce consistent, stable, error-characterized global BA information. The project includes as well developing algorithms to generate georeferenced and calibrated reflectances of (A)ATSR, VEGETATION and MERIS data, identifying potential sources of confusion with burned areas (clouds, smoke, cloud shadows, water, snow, topographic shadows). The final product will be a merging of BA information derived from three different sensors . The outputs will be adapted to the needs of the atmospheric and vegetation modelling communities.

International audience; An emerging way to reduce the geodetic parameter uncertainty is to combine the large numbers of data provided by satellite radar images. However, the measurements by radar images are subjected both to random and epistemic uncertainties. Thus, mathematical theories which are adequate for each type of uncertainty representation and handling have to be selected. Probability theory is known as the adequate theory for uncertainties corresponding to random variables, but questionable for epistemic uncertainties, arising from information incompleteness. Fuzzy theory being a generalization of interval mathematics, it is more adapted to such uncertainty. Moreover it provides a bridge with probability theory by its ability to represent a family of probability distributions. Therefore, we consider here the conventional probability and the fuzzy approaches for handling the random and epistemic uncertainties of D-InSAR and SAR correlation measurements. The applications are performed on the measurement of displacement field due to the Kashmir earthquake (2005).

18 pages; The challenging task of biomass prediction in dense and heterogeneous tropical forest requires a multi-parameter and multi-scale characterization of forest canopies. Completely different forest structures may indeed present similar above ground biomass (AGB) values. This is probably one of the reasons explaining why tropical AGB still resists accurate mapping through remote sensing techniques. There is a clear need to combine optical and radar remote sensing to benefit from their complementary responses to forest characteristics. Radar and Lidar signals are rightly considered to provide adequate measurements of forest structure because of their capability of penetrating and interacting with all the vegetation strata. However, signal saturation at the lowest radar frequencies is observed at the midlevel of biomass range in tropical forests (Mougin et al. 1999; Imhoff, 1995). Polarimetric Interferometric (PolInsar) data could improve the inversion algorithm by injecting forest interferometric height into the inversion of P-band HV polarization signal. Within this framework, the TROPISAR mission, supported by the Centre National d'Etudes Spatiales (CNES) for the preparation of the European Space Agency (ESA) BIOMASS program is illustrative of both the importance of interdisciplinary research associating forest ecologists and physicists and the importance of combined measurements of forest properties. Lidar data is a useful technique to characterize the vertical profile of the vegetation cover (e.g. Zhao et al. 2009) which in combination with radar (Englhart et al. 2011) or optical (e.g. Baccini et al. 2008; Asner et al. 2011) and field plot data may allow vegetation carbon stocks to be mapped over large areas of tropical forest at different resolution scales ranging from 1 hectare to 1 km². However, small-footprint Lidar data are not yet accessible over sufficient extents and with sufficient revisiting time because its operational use for tropical studies remains expensive. At the opposite, very-high (VHR) resolution imagery, i.e. approximately 1-m resolution, provided by recent satellite like Geoeye, Ikonos, Orbview or Quickbird as well as the forthcoming Pleiades becomes widely available at affordable costs, or even for free in certain regions of the world through Google Earth®. Compared to coarser resolution imagery with pixel size greater than 4 meters, VHR imagery greatly improves thematic information on forest canopies. Indeed, the contrast between sunlit and shadowed trees crowns as visible on such images (Fig. 1) is potentially informative on the structure of the forest canopy while new promising methods now exist for analyzing these fine scale satellite observations (e.g. Bruniquel-Pinel & Gastellu-Etchegorry, 1998; Malhi & Roman-Cuesta, 2008; Rich et al. 2010). Besides, we believe that there is also a great potential in similarly using historical series of digitized aerial photographs that proved to be useful in the past for mapping large extents of unexplored forest (Le Touzey, 1968; Richards, 1996) for quantifying AGB changes through time. This book chapter presents the advancement of a research program undertaken by our team for estimating high biomass mangrove and terra firme forests of Amazonia using canopy grain from VHR images (Couteron et al. 2005; Proisy et al. 2007; Barbier et al., 2010; 2011). We present in a first section, the canopy grain notion and the fundamentals of the Fourier-based Textural Ordination (FOTO) method we developed. We then introduce a dual experimental-theoretical approach implemented to understand how canopy structure modifies the reflectance signal and produces a given texture. We discuss, for example, the influence of varying sun-view acquisition conditions on canopy grain characteristics. A second section assesses the potential and limits of the canopy grain approach to predict forest stand structure and more specifically above ground biomass. Perspectives for a better understanding of canopy grain-AGB relationships conclude this work.

Coastal ocean-color estimation needs to retrieve not only molecular and aerosol scattering ( ρ a ), but also high spatial resolution sea-surface reflectance ( ρ a ) because ρ g has fine temporal and spatial scales due to variable winds and air-sea stability caused by the coastal geographical structure. Murakami and Frouin 2008 showed a possibility of ρ g correction by using near infrared (NIR) and shortwave infrared (SWIR) channels of MODIS 500m observations. This study investigated the correction of the atmospheric and sea-surface reflectance on the southwest of New Caledonia lagoon using AVNIR-2 which has 10-m resolution but doesn't have SWIR. After corrections of gas absorption and molecule scattering, we estimated ρ a + ρ g and water-leaving reflectance iteratively through IOPs retrieved from visible bands. Spectral slope of ρ a + ρ g was assumed uniform within our small target area (60km×40km). We tested sensitivity to several possible IOP spectra (total absorption of particle and dissolved matter and back-scattering coefficients) with comparison to in-situ IOP measurements. The AVNIR-2 derived remote sensing reflectance agreed well to the MODIS one (rootmean square difference / average of Rrs 443nm was 43%), and AVNIR-2 IOPs agreed well to in-situ IOP measurements (correlation coefficients more than 0.9) when we used the IOP spectra modeled by in-situ measurements around the New Caledonia. Chlorophyll-a ( Chla ) calculated by the AVNIR-2 IOPs showed better agreement to in-situ Chla in the lagoon areas where traditional blue/green algorithms overestimated.

No map of the sea floor is available yet on the whole lagoon of New Caledonia. We tried to validate a method to map it with MeRIS images on the south western part of the lagoon. The non-linear effect of water column light attenuation can then be corrected to obtain the absolute reflectance of the seabed. Light attenuation by the water column can be determined by comparing the radiance of standard features on the seabed at different depth. Bathymetry can also be determined by measuring the relative reflectance of the seabed in green and red light spectral bands. Once the effect of attenuation has been removed, a supervised classification can be applied in order to obtain the location of each item on the sea floor. Validations are operated with ground measurements of depth, spectral profiles and some available maps.

With the objective of budding climatological statistics on propagating African convective systems and extracting pertinent parameters for the evaluation of precipitation an automatic method of tracking clouds with cold tops on Meteosat infrared (IR) images is established. This method takes into account eventual separation or merging of clouds. Results obtained in several cases not only demonstrate the capability of the method to perform correct tracking in different situations, but also show that objective determination of parameters, such as propagation speeds, system area, and volume index, is possible. The analysis of the time evolution of these last parameters allows a clear characterization of the cloud life cycle with its growing and decreasing stages, which may be useful for improving precipitation-estimation methods based only on cold-cloud occurrences or cloud-top temperatures.

International audience; This study describes a comprehensive method to produce routinely regional maps of seasonal snow cover in the Southern Alps of New Zealand (upper Waitaki basin) on a subpixel basis, and with the MODerate Resolution Imaging Spectroradiometer (MODIS). The method uses an image fusion algorithm to produce snow maps at an improved 250 m spatial resolution in addition to the 500 m resolution snow maps. An iterative approach is used to correct imagery for both atmospheric and topographic effects using daily observations of atmospheric parameters. The computation of ground spectral reflectance enabled the use of image-independent end-members in a constrained linear unmixing technique to achieve a robust estimation of subpixel snow fractions. The accuracy of the snow maps and performance of the algorithm were assessed carefully using eight pairs of synchronic MODIS/ASTER images. ‘Pixel-based' metrics showed that subpixel snow fractions were retrieved with a Mean Absolute Error of 6.8% at 250 m spatial resolution and 5.1% after aggregation at 500 m spatial resolution. In addition, a ‘feature-based' metric showed that 90% of the snowlines were depicted generally within 300 m and 200 m of their correct position for the 500-m and 250-m spatial resolution snow maps, respectively. A dataset of 679 maps of subpixel snow fraction was produced for the period from February 2000 to May 2007. These repeated observations of the seasonal snow cover will benefit the ongoing effort to model snowmelt runoff in the region and to improve the estimation and management of water resources.

Remote sensing makes it possible to quantify the extent of the clearings in the Brazilian Amazon. The diversity of sensors also enables a multi-scale approach to the landscapes of Mato Grosso, while pinpointing the respective role of the actors in deforestation: large stockbreeders, small settlers, and farmers integrated into the soy, cotton or sugarcane agro-business. The landscapes, originally very different, are moving towards a single type where forest occupies only a tiny area (reserves, ancillary forests) in the middle of pasture or cultivated fields., La télédétection permet de suivre dans le temps et de quantifier l'ampleur des défrichements en Amazonie brésilienne. La diversité des capteurs permet aussi une approche multiscalaire des paysages du Mato Grosso tout en précisant le rôle respectif des acteurs du défrichement : grands propriétaires éleveurs, petits colons, ou cultivateurs intégrés à l'agrobusiness du soja, du coton ou de la canne à sucre. Les paysages, très différents à l'origine, convergent vers un type unique où la forêt n'occupe plus qu'un maigre espace (réserves légales, forêts-galeries) au milieu des pâturages ou des champs cultivés.

We propose to implement the ARSIS concept to fuse the high spatial content of the two 250 m spectral bands of MODIS into its five 500 m bands using wavelet based multi- resolution analysis. Our objective was to test the effectiveness of this technique to increase the accuracy of snow mapping in mountainous environments. To assess the performance of this approach, we took advantage of the simultaneity between ASTER and MODIS sensors, both on the TERRA platform. With a 15 m spatial resolution, the ASTER sensor provided reference snow maps that were then compared to MODIS derived snow maps. The benefit of the method was assessed through the investigation of various metrics. The enhanced snow map are therefore of great benefit for environmental and hydrological applications in steep terrain.

This paper presents the results of object-based techniques applied to mapping and hi-temporal change detection of mangrove ecosystem in Low Casamance, Senegal. The methodology was applied on SPOT images for 1986 and 2006. Reference data was based on ground truth and visual verification of very high resolution images. High accuracy was obtained in mapping of mangrove cover. Other land cover classes showed however lower accuracy. The segmentation process for the change detection analysis was executed on a multi-date image and classification was then completed with a standard nearest neighbour classifier. Change detection analysis in mangrove ecosystems proved to be particularly difficult due to underlying presence of water and tide influences. Unlikely most mangrove ecosystems worldwide, important negative trends were not identified and low dynamicity was recorded during the two dates. This information questions the generalized preconception over mangrove retreat. The methodology will be further extended to Sine-Saloum the other main mangrove ecosystem of Senegal and should provide updated mangrove inventory for the country.

Monitoring structural organization and biomass of tropical forest from remote sensing observations constitutes a great challenge. For this, texture analysis of very high spatial resolution (VHR) data can significantly contribute to understand the link between image aspect and forest structure. Within this context, our recent works argued that swapping the spatial domain for the frequency domain permits to consistently characterize vegetation and landscape spatial patterns through coarseness/fineness textural indices. In this paper, we synthesize the main results obtained on the characterization of canopy grain using Fourier-based textural ordination (FOTO) of very high resolution images acquired over terra firme rain forest and high biomass mangroves. In addition, we compare texture indices originating from aerial photographs and Ikonos images. Requirements for developing a standardized measurement of forest canopy grain with broad regional validity are finally discussed.

DOI:10.1016/j.jhydrol.2005.03.027; The knowledge of evapotranspiration is of great interest in many applications such as assessing irrigation water use. In this paper, a methodology is presented to estimate evapotranspiration using the surface energy balance model S-SEBI (Simplified Surface Energy Balance Index) and the evaporative fraction. The method is applicable under the assumptions of constant atmospheric conditions and sufficient wet and dry pixels over the image. The model uses remotely sensed parameters such as albedo, MSAVI (Modified Soil Adjusted Vegetation Index) and surface temperature and emissivity images. The methodology has been applied over the Barrax test site, located in the province of Albacete (Castilla La Mancha, Spain), in the framework of the DAISEX (Digital Airborne Imaging Spectrometer Experiment) campaigns carried out by ESA (European Space Agency). Net radiation, soil heat fluxes and evapotranspiration have been obtained from DAIS images for alfalfa, corn and grass plots. The results show that alfalfa plot gives the highest values, around 5 mm dK1, whereas corn plot gives the lowest, around 3.5 mm dK1. The proposed methodology leading to an accuracy for the daily evapotranspiration lower than 1 mm dK1.

A neural network is developed to retrieve chlorophyll a concentration from marine reflectance by use of the five visible spectral bands of the Sea-viewing Wide Field-of-view Sensor (SeaWiFS). The network, dedicated to the western equatorial Pacific Ocean, is calibrated with synthetic data that vary in terms of atmospheric content, solar zenith angle, and secondary pigments. Pigment variability is based on in situ data collected in the study region and is introduced through nonlinear modeling of phytoplankton absorption as a function of chlorophyll a, b, and c and photosynthetic and photoprotectant carotenoids. Tests performed on simulated yet realistic data show that chlorophyll a retrievals are substantially improved by use of the neural network instead of classical algorithms, which are sensitive to spectrally uncorrelated effects. The methodology is general, i.e., is applicable to regions other than the western equatorial Pacific Ocean.

The capacity of the Landsat 7 Enhanced Thematic Mapper Plus sensor to classify the shallow benthic ecosytems of New Caledonia (South Pacific) is tested using a novel unsupervised classification method. The classes are defined by using a set of multiple spectral decision rules based on the image spectral bands. A general model is applied to the entire Southwest lagoon (5500 km(2)) and tested on three representative sites: a section of the barrier reef, a cay reef flat rich in corals, and a cay reef flat rich in algae and seagrass beds. In the latter one, the classification results are compared with a locally optimized model, with aerial color photographs and extensive ground-truthed observations. Results show that a reconnaissance of the main benthic habitats in shallow areas (, Le but de cette étude est de tester la capacité des images Landsat 7 Enhanced Thematic Mapper Plus à discriminer les principales classes d’habitats benthiques rencontrées dans les parties peu profondes du système récifal et lagonaire de Nouvelle-Calédonie (Pacifique Sud). Une méthode originale de classification non-supervisée est proposée. Les habitats benthiques correspondent à une combinaison de plusieurs règles de décision établies à partir des bandes radiométriques Landsat. Cette modélisation statistique des habitats benthiques est appliquée au lagon sud-ouest de Nouvelle-Calédonie (5500 km2). Les résultats sont testés sur trois sites témoins contrastés: un platier de récif barrière, un platier d’îlot riche en corail et un platier d’îlot riche en herbiers/algueraies. Pour ce dernier, le résultat est comparé à celui d’un modèle optimisé, construit à échelle locale et validé à partir de photographies aériennes et d’observations de terrain. Les résultats montrent qu’une reconnaissance des différentes classes benthiques est possible pour des fonds peu profonds (< 5 m de profondeur), à l’échelle géomorphologique pour les structures récifales et à l’échelle des habitats pour les herbiers. Toutefois, les résultats bruts du modèle doivent être interprétés en fonction de critères spatiaux pour corriger les confusions entre certaines classes, notamment entre les pentes coralliennes et les herbiers denses.

Cette étude a pour objet l’extraction d’information sur les pistes forestières en milieu tropical par l’utilisation de l’imagerie satellitaire. Un procédé simple a été mis au point afin d’analyser des images entières de façon aisée et rapide. Le résultat, qui se veut robuste, consiste en la réalisation d’un plan vecteur intégrable dans un système d’information géographique. L’actualisation régulière de ce vecteur permet alors de contrôler l’extension du réseau et de l’impact de l’exploitation sur le massif forestier., The objective of this study was to highlight logging tracks in tropical forests using remote sensing techniques. A simple processing system was built up to analyse images easily and quickly. This has produced a robust result consisting of a vector layer in a format which is compatible with geographical information systems. By updating this vector layer, the logging track network and logging impacts can be monitored., La finalidad de este estudio es la obtención de información sobre las pistas forestales en medio tropical mediante la utilización de imágenes satelitarias. Se puso a punto un método sencillo para analizar imágenes enteras de manera simple y rápida. El resultado, que intenta ser fiable, consiste en la realización de un plan vector integrable en un sistema de información geográfica. La actualización regular de este vector permite controlar la extensión de la red y el impacto de la explotación en el macizo forestal.

Tunas are known to be able to travel long distances. The aim of this paper is to propose new ethological models which reproduce some tuna movements using the dynamics of their environment. We use sea surface temperature animations (from remote sensing data) to model the South West Indian Ocean, and French purse seiners data are used to estimate movements of fish. The objective of the models will be to find a northern movement from the Mozambique Channel to the Seychelles Islands at the appropriate time (May–July). The initial model uses our ecological knowledge of tunas, i.e. the search behavior for high concentrations of food commonly associated with thermal fronts. In some cases, this simple model creates some northern movements from the Mozambique Channel, but it cannot be used to reproduce large-scale movements between the Mozambique Channel and the Seychelles Islands. The next generation model is created where tuna behaviors are modeled by an artificial neural network, using a genetic algorithm to adjust the connection weights. The tuna school-network receives daily information from its local environment and chooses the best actions in order to be able to pass from the Mozambique Channel to the Seychelles Islands at the appropriate time. One neural network emerges and represents an adaptive behavior able to interpret daily sea surface temperatures to mimic large-scale tuna movements. This artificial behavior can be generalized to each possible departure position from the Mozambique Channel. This modelling represents a new tool to study large-scale movements of pelagic fish, and is a first step towards real-time management of fisheries.

Toward a classification of landscape units in a mountain region (Central Nepal). — Land occupance in mountainous regions depends in part on slopes exposure and climatic zonation. Spectral analysis of landscape units also reveals variations which are due to the level of illumination on the slope at the time when the satellite images were actually registered. There are special treatments now which enable one to eliminate shadow effects which are due to topography; themes may now be treated separately in terms of the degree of illumination. The identification of these themes and their different light settings should follow an automatic classification system based on radiometric and other measures, e.g., parameters of texture and environment. This clustering procedure is based on analysis of variance in the occurrence and diversity of the major radiometric values found within a given grid square (spectral density). Within a given altitudinal zone, too, correlations can be calculated among spectral responses within a given family of identical themes, in different lighting contexts. Accurate information on «ground-truth» is found from a cartographic survey of the experimental region and through a supervised analysis of the images themselves (vegetation indices, parameters of texture, multi-spectral and multi-temporal analysis)., Dans les régions montagneuses, l'occupation des sols dépend entre autres de l'exposition des versants et de l'étagement des climats. Les réponses spectrales des unités de paysage y varient également en fonction de l'éclairement des pentes au moment des prises de vue satellitaires ; on peut tenter d'éliminer par des traitements appropriés les effets d'ombre dus à la topographie, et traiter séparément les thèmes selon la nature de l'éclairement : leur identification dans des plages à éclairement différent devra suivre un processus de classification automatique dépendant de la radiométrie et de ses dérivés, tels que paramètres de texture et d'environnement. Ce seuillage repose sur l'élaboration de variables dépendant de la variance et de l'occurrence des valeurs radiométriques majoritaires comprises dans une maille carrée (compacité radiométrique). On cherche ensuite à établir, pour une même tranche altitudinale, les corrélations existant éventuellement entre les réponses spectrales d'une famille de thèmes identiques (paysage) placés dans des conditions d'éclairement différent. La connaissance de la vérité de terrain est acquise grâce à un relevé cartographique dans les régions témoins et à un traitement supervisé des images (indices de végétation et paramètres de texture, analyse multibande et multitemporelle)., Blamont Denis,Mering Catherine,Parrot Jean-FrançoisBlamont Denis, Mering Catherine, Parrot Jean-François. Essai de classification des unités de paysage en région montagneuse (centre du Népal). In: L'Espace géographique, tome 13, n°3, 1984. pp. 225-232.