Institute and Observatory of Geophysics of Antananarivo

Abstract East African Rift System plate geometries and surface motions are some of the least constrained in the context of global plate motion models. In this study, we used GPS data to constrain Somalian plate rotation and to suggest a new tectonic plate geometry for the region. In addition, we tested geologic data from the Southwest Indian Ridge and new GPS data on Madagascar to determine refined kinematics of the Lwandle microplate. A zone of broad deformation was discovered, extending from the eastern boundary of the Rovuma microplate, across the Comoros Islands, and including parts of central and northern Madagascar. Madagascar is fragmenting, with southern Madagascar rotating with the Lwandle microplate and a piece of eastern and south-central Madagascar moving with the Somalian plate. Divergence of the Nubian-Somalian plate system across the East African Rift System involves both diffuse deformation and strain accommodation along narrow rift segments that bound rigid blocks.

Abstract. The Congo River basin (CRB) is the second largest river system in the world, but its hydroclimatic characteristics remain relatively poorly known. Here, we jointly analyse a large record of in situ and satellite-derived observations, including a long-term time series of surface water height (SWH) from radar altimetry (a total of 2311 virtual stations) and surface water extent (SWE) from a multi-satellite technique, to characterize the CRB surface hydrology and its variability. First, we show that SWH from altimetry multi-missions agrees well with in situ water stage at various locations, with the root mean square deviation varying from 10 cm (with Sentinel-3A) to 75 cm (with European Remote Sensing satellite-2). SWE variability from multi-satellite observations also shows a plausible behaviour over a ∼25-year period when evaluated against in situ observations from the subbasin to basin scale. Both datasets help to better characterize the large spatial and temporal variability in hydrological patterns across the basin, with SWH exhibiting an annual amplitude of more than 5 m in the northern subbasins, while the Congo River main stream and Cuvette Centrale tributaries vary in smaller proportions (1.5 to 4.5 m). Furthermore, SWH and SWE help illustrate the spatial distribution and different timings of the CRB annual flood dynamic and how each subbasin and tributary contribute to the hydrological regime at the outlet of the basin (the Brazzaville/Kinshasa station), including its peculiar bimodal pattern. Across the basin, we estimate the time lag and water travel time to reach the Brazzaville/Kinshasa station to range from 0–1 month in its vicinity in downstream parts of the basin and up to 3 months in remote areas and small tributaries. Northern subbasins and the central Congo region contribute highly to the large peak in December–January, while the southern part of the basin supplies water to both hydrological peaks, in particular to the moderate one in April–May. The results are supported using in situ observations at several locations in the basin. Our results contribute to a better characterization of the hydrological variability in the CRB and represent an unprecedented source of information for hydrological modelling and to study hydrological processes over the region.

Abstract. Within the framework of the UV-Indien network, nine ground stations have been equipped with ultraviolet broadband radiometers, five of them have also been equipped with an all-sky camera, and the main station in Saint-Denis de la Réunion is also equipped with a spectroradiometer. These stations are spatially distributed to cover a wide range of latitudes, longitudes, altitudes, and environmental conditions in five countries of the western Indian Ocean region (Comoros, France, Madagascar, Mauritius, and Seychelles), a part of the world where almost no measurements have been made so far. The distribution of the stations is based on the scientific interest of studying ultraviolet radiation not only in relation to atmospheric processes but also in order to provide data relevant to fields such as biology, health (prevention of skin cancer), and agriculture. The main scientific objectives of this network are to study the annual and inter-annual variability in the ultraviolet (UV) radiation in this area, to validate the output of numerical models and satellite estimates of ground-based UV measurements, and to monitor UV radiation in the context of climate change and projected ozone depletion in this region. A calibration procedure including three types of calibrations responding to the various constraints of sustaining the network has been put in place, and a data processing chain has been set up to control the quality and the format of the files sent to the various data centres. A method of clear-sky filtering of the data is also applied. Here, we present an intercomparison with other datasets, as well as several daily or monthly representations of the UV index (UVI) and cloud fraction data, to discuss the quality of the data and their range of values for the older stations (Antananarivo, Anse Quitor, Mahé, and Saint-Denis). Ground-based measurements of the UVI are used to validate satellite estimates – Ozone Monitoring Instrument (OMI), the TROPOspheric Monitoring Instrument (TROPOMI), and the Global Ozone Monitoring Experiment (GOME) – and model forecasts of UVI – Tropospheric Emission Monitoring Internet Service (TEMIS) and Copernicus Atmospheric Monitoring Service (CAMS). The median relative differences between satellite or model estimates and ground-based measurements of clear-sky UVI range between −34.5 % and 15.8 %. Under clear skies, the smallest UVI median difference between the satellite or model estimates and the measurements made by ground-based instruments is found to be 0.02 (TROPOMI), 0.04 (OMI), −0.1 (CAMS), and −0.4 (CAMS) at Saint-Denis, Antananarivo, Anse Quitor, and Mahé, respectively. The diurnal variability in UVI and cloud fraction, as well as the monthly variability in UVI, is evaluated to ensure the quality of the dataset. The data used in this study are available at https://doi.org/10.5281/zenodo.4811488 (Lamy and Portafaix, 2021a).

In Madagascar, allometric models have been developed only for the primary forests. There have been no specific models for the secondary forests. However, with a view to extending the geographic scope of the REDD+1 mechanism, it deems important to consider the secondary forests of which the potential of carbon sequestration remains unknown. In eastern part of Madagascar, the secondary forests are mostly colonized by Ravenala madagascariensis, an endemic monocotyledonous plant. The main goal of this study was to develop local allometric models for secondary forests with Ravenala madagascariensis. Models for Ravenala madagascariensis were developed separately from those of other woody species. The second goal was to test the validity of the existing allometric equations for the woody species into the study sites, namely three pantropical models developed by Brown, Chave et al. and the best fitted local model developed by Vieilledent et al. for the dense humid forest in Madagascar. Using the destructive method, 120 stems of Ravenala madagascariensis and 115 trees were harvested, weighed and measured. As results, the best fitted local model for the woody species was AGB=EXP(−1.515+1.912ln(D)+0.471ln(H)+0.732ln(ρ)), where AGB is the aboveground biomass in kg, H the total height in m, D the diameter at the breast height in cm and ρ the wood specific gravity. For Ravenala madagascariensis, the model which only considered total height was the most appropriate: AGB=EXP(−4.996+5.654ln(H)−0.772(ln(H))2). Comparing with the existing equations, errors of estimates were higher for the pantropical models and a slight overestimation was observed while applying that of Vieilledent et al. It is thus recommended to choose the local site specific models for estimating the aboveground biomass for the secondary forests with Ravenala madagascariensis in the eastern part of Madagascar.

The understanding of physical and mechanical properties of rock is considered as critical in drilling, geo-engineering, and construction applications. As an example, the awareness of these rock parameters contributes to avoid or minimizing instability around the wellbore while drilling. The laboratory experiment of understanding of these parameters can be done in two-different ways: static, where the sample subjects to destruction after the test and dynamic, known as non-destruction method. The non-destructive method using ultrasonic waves under a series of different stress conditions, starting from 7 to 56 MPa with incrementation of 7MPa, has been used in this paper in order to characterize the mechanical properties of dry Zbylutów sandstone at 20 and 80°C. The velocity of primary (P) and secondary (S) waves within these ranges has been recorded in order to understand the behavior of the mechanical properties. The results showed that the Young’s modulus, bulk modulus, shear modulus, and Lame’s constant of Zbylutów sandstone have a positive correlation with good coefficient correlation with the increased stress, while the Poisson’s ratio showed a negative correlation. Besides, the effect of temperature on the rock parameters is approved by the decrease of primary wave velocity in this two-different temperature range. Such results are necessary when preparing the appropriate mud weight for drilling process, which is related to wellbore instability.

This paper presents an operational approach for detecting floods and establishing flood extent using Sentinel-1 radar imagery with Google Earth Engine. Flooded areas are identified using a change-detection method based on the normalized difference. The HAND algorithm is used to delineate zones for processing. The approach was tested and calibrated at small scale to identify optimal parameters for flood detection. It was then applied to the whole of the island of Madagascar after the cyclone Batsirai in 2022. The proposed method is enabled by the computing power and data availability of Google Earth Engine and Google Colab. The results show satisfactory accuracy in delineating flooded areas. The advantages of this approach are its rapidity, online availability and ability to detect floods over a wide area. The approach relying on Google tools thus offers an effective solution for generating a large-scale synoptic picture to inform hazard management decision-making. However, one of the method’s drawbacks is that it depends to a large extent on frequent radar imagery being available at the time of flood events and on free access to the platform. These drawbacks will need to be taken into account in an operational scenario.

Les marais potentiellement aménageables en rizières constituent des zones de future défriche en zones forestières tropi\-cales. Une chaîne de traitement d'images satellitaires multirésolutions et multisources, utilisant Orfeo ToolBox, est mise à l'épreuve pour discriminer les zones humides : eaux, marais et rizières. Cette méthodologie combine des indices radiométriques extraits d'images qui disposent d'un plus grand nombre de bandes spectrales et des indices texturaux issus d'images à résolution spatiale élevée. Ainsi l'information spectrale d'une image du satellite Landsat~7 ETM+ est valorisée pour identifier les zones humides avec une résolution de 30$\:$m et en dresser la cartographie régionale. L'information texturale d'images du satellite SPOT~5 de 10 m et de 2,5$\:$m de résolution est utilisée pour discriminer des types de zones humides et les cartographier à l'échelle locale. En combinant les données des deux satellites, les surfaces en eaux, marais et rizières ont été évaluées avec un indice de Kappa égal à 0,8, dans deux communes du corridor forestier de Fianarantsoa (Madagascar).

Cette publication participe à la compréhension de la dégradation de l’environnement à Madagascar. Spécifiquement, il s’agit d’analyser la distribution spatiale et temporelle des feux à Madagascar et d’identifier les facteurs à l’origine des feux. Les données fournies par le capteur MODIS sur l’ensemble du territoire de Madagascar pour les années 2014 à 2018 sont utilisées. Les résultats obtenus indiquent une certaine stabilité des points de feux actifs observés à Madagascar pendant la période 2014 à 2018, sauf pour l’année 2016, année de forte sécheresse mondiale. Les feux brûlent en moyenne une superficie de 3 757 792 ha par an. La majorité de ces feux est observée pendant la saison sèche et leur maximum en nombre et en superficie est atteint aux mois de septembre et octobre. Au premier abord, les feux couvrent la quasi-totalité du pays. Toutefois, les analyses spatiales et temporelles montrent que les points de feu sont regroupés en agrégats d’ordre 1, 2, 3 ou 4. Les agrégats indiquent les zones où l’on retrouve une concentration de points de feux. Les régions contenant des agrégats d'ordres supérieurs sont les plus exposées. Ce sont des zones auxquelles il est nécessaire d’accorder une attention particulière. Les résultats de l’analyse par permutation spatio-temporelle indiquent que les feux proches dans l’espace le sont aussi dans le temps. La répartition des agrégats par rapport aux pistes et aux villages démontre que les feux se propagent essentiellement à proximité des lieux habités et des voies d’accès.

ABSTRACT Land degradation contributes to loss of land productivity, climate change, and biodiversity loss worldwide, making the monitoring of land cover change a fundamental component of sustainable land use planning. To facilitate this task, international research organizations publish increasingly high‐quality global land cover data, yet those are still proving too imprecise at the scale of Small Island Developing States (SIDS) such as Mauritius. Though rapidly developing, Mauritius does not have a locally produced land cover map which is consistently updated and made available to the public in the scope of participatory planning processes. This study sought to bridge this gap by revising a 2010 land cover map of Mauritius island and producing a new map for 2020, which can serve as a basis for future updates through a consistent methodology that combines photointerpretation and semi‐automated classification from various datasets. These maps enabled the observation of key changes in the island's landscape, which are consistent with socio‐economic events, namely the decline of agricultural coverage (12.7%) due to the phasing out of preferential trade agreements for sugar, and the expansion of real estate (14.3% increase in built area) following new investment promotion policies. Considering the small size and high population density of the island, as is the case for SIDS, these developments and the Land Degradation Neutrality Targets of Mauritius bring the feasibility of area‐based land degradation neutrality into question and call for the analysis of land and ecosystem conditions through additional indicators.

SUMMARY The lithosphere of Madagascar records a long series of tectonic processes. Structures initially inherited from the Pan-African Orogeny are overprinted by a series of extensional tectonic and magmatic events that began with the breakup of Gondwana and continued through to the present. Here, we present a Pn-tomography study in which Pn traveltimes are inverted to investigate the lateral variation of the seismic velocity and anisotropy within the uppermost mantle beneath Madagascar. Results show that the Pn velocities within the uppermost mantle vary by ±0.30 km s–1 about a mean of 8.10 km s–1. Low-Pn-velocity zones (<8.00 km s–1) are observed beneath the Cenozoic alkaline volcanic provinces in the northern and central regions. They correspond to thermally perturbed zones, where temperatures are estimated to be elevated by ∼100–300 K. Moderately low Pn velocities are found near the southern volcanic province and along an E–W belt in central Madagascar. This belt is located at the edge of a broader low S-velocity anomaly in the mantle imaged in a recent surface wave tomographic study. High-Pn-velocity zones (>8.20 km s–1) coincide with stable and less seismically active regions. The pattern of Pn anisotropy is very complex, with small-scale variations in both the amplitude and the fast-axis direction, and generally reflects the complicated tectonic history of Madagascar. Pn anisotropy and shear wave (SKS) splitting measurements show good correlations in the southern parts of Madagascar, indicating coherency in the vertical distribution of lithospheric deformation along Pan-African shear zone as well as coupling between the crust and mantle when the shear zones were active. In most other regions, discrepancies between Pn anisotropy and SKS measurements suggest that the seismic anisotropy in the uppermost mantle beneath Madagascar differs from the vertically integrated upper mantle anisotropy, implying a present-day vertical partitioning of the deformation. Pn anisotropy directions lack the coherent pattern expected for an incipient plate boundary within Madagascar proposed in some kinematic models of the region.

Abstract. To gain a deeper understanding of cloud variability over the Southwest Indian Ocean (SWIO) region, various measurement techniques can be used. Cloud data focused on two main sites (Antananarivo, Madagascar, and Reunion Island) were collected over nearly three years (September 2019 to June 2022) using ground-based all-sky camera and Meteosat Second Generation (MSG) satellite imagery. This study primarily examined cloud fraction, although additional analysis based on cloud types is also needed. Two image processing algorithms were used to estimate cloud fraction by analyzing the camera images: one provided by the manufacturer Reuniwatt, and the other, Elifan, initially developed by CNRS. Their performance was compared to assess their relative strengths. Additionally, comparisons with MSG satellite data were carried out to ensure consistency and evaluate the complementarity of ground-based and satellite observations. Despite differences in methodology, the various data sources showed strong agreement. The results indicate that in Antananarivo, during the dry season, a high morning cloud fraction (∼50 %) is typically observed, which gradually decreases throughout the day. In the wet season (December to April), cloud fraction varies between approximately 30 % and 60 %, with reduced cloudiness observed around midday in October and November. In contrast, in Saint-Denis, Reunion, skies are generally clear in the morning but become increasingly cloudy as the day progresses, reaching up to 80 % cloud fraction during the wet season and around 60 % during the dry season.

La méthode gravimétrique a été utilisée dans le bassin sédimentaire de la zone semi-aride du Sud de Madagascar pour déterminer la morphologie du toit du substratum, qui est constitué de roches cristallines et volcaniques. La structure morphologique du substratum influe sur l'hydrogéologie et la dynamique de l'eau souterraine de la zone. La corrélation entre la répartition de la conductivité électrique de l'eau et la morphostructure du substratum (forme et linéament) confirme l'étroite dépendance des deux paramètres, et permet par la suite de délimiter la zone favorable à l'exploitation future de l'eau.

Noise due to surface wind and temperature is a problem in infrasound. Efficiency of IMS network concerns scientists. It is obvious to find the causes of deficiencies of detection of infrasound station by studying background noise power with respect to the surface wind and the temperature. Data measured by MB2000 microbarometer of infrasound station I33MG are used for the study. Infrasound records are separated into 4 frequency bands centered respectively at: 1 Hz, 0.25 Hz, 0.0625 Hz and 0.0156 Hz. Effects of surface wind and temperature are studied by plotting the variations of the background noise power with respect to the temperature or wind speed in the four considered frequency bands and compared with the median of background noise power. The influence of temperature is manifested by a reduction in the number of low-frequency detection. The surface wind reduces the number of detection at a high frequency. An exponential function is proposed to predict the variations of the noise power in different observation frequencies and temperature and wind conditions. The views expressed herein are those of the authors and do not necessarily reflect the views of the CTBTO Preparatory Commission.

Since October 2016, 10 ground-based stations have been progressively equipped for continuous measurements of UV indices and atmospheric cloud fraction. All these instruments form the UV-Indien measurement network. The stations are homogeneously distributed in 5 countries of the Western Indian Ocean region (Comoros, France, Madagascar, Mauritius and Seychelles). The main scientific objectives of this network are to study the annual and inter-annual variability of the ultraviolet (UV) radiation in this area, to validate the output of numerical models and satellite estimates of ground-based UV measurements, and to monitor UV radiation in the context of climate change and projected ozone depletion in this region.

Abstract. As part of the UV-Indien Network, 9 ground-based stations have been equipped with one spectroradiometer, radiometers and all-sky cameras. These stations are homogeneously distributed in 5 countries of the Western Indian Ocean region (Comoros, France, Madagascar, Mauritius and Seychelles), a part of the world where almost no measurements have been made so far. The main scientific objectives of this network are to study the annual and inter-annual variability of the ultraviolet (UV) radiation in this area, to validate the output of numerical models and satellite estimates of ground-based UV measurements, and to monitor UV radiation in the context of climate change and projected ozone depletion in this region. The first results are presented here for the oldest stations (Antananarivo, Anse Quitor, Mahé and Saint-Denis). Ground-based measurements of UV index (UVI) are compared against satellite estimates (Ozone Monitoring Instrument (OMI), the TROPOspheric Monitoring Instrument (TROPOMI), the Global Ozone Monitoring Experiment (GOME) and model forecasts of UVI (Tropospheric Emission Monitoring Internet Service (TEMIS) and Copernicus Atmospheric Monitoring Service (CAMS). The median relative differences between satellite or model estimates and ground-based measurements of clear-sky UVI range between −34.5 % and 15.8 %. Under clear skies, the smallest UVI median difference between the satellites or model estimates and the measurements of ground-based instruments is found to be 0.02 (TROPOMI), 0.04 (OMI), −0.1 (CAMS) and −0.4 (CAMS) at St-Denis, Antananarivo, Anse Quitor and Mahé respectively. The cloud fraction and UVI diurnal profile are calculated for these four stations. The mean UVI values at local solar noon range between 10 (Antananarivo, Anse Quitor and Saint-Denis) and 14 at Mahé. The mean UVIs in clear-sky conditions are higher than mean UVI in all-sky conditions, although it can still be noted that UVI maxima are higher for all-sky conditions than for clear sky conditions. This is the result of UVI enhancement induced by clouds, observed at these four stations. The greatest increase in UV radiation under cloudy conditions was observed at the Mahé station, with increases of more than 4. The data used in this study is available at https://doi.org/10.5281/zenodo.4572026 (Lamy and Portafaix, 2021).

<p>Radiation (UV) is one of the main components of solar radiation transmitted by the Earth's atmosphere. Exposure to UV radiation can have both positive and negative effects on the biosphere and humans in particular. Overexposure significantly increases the risk of skin cancer and eye problems.</p><p>Ozone, cloud cover and zenithal solar angle are the main parameters affecting UV radiation levels at the surface. Stratospheric ozone in particular strongly absorbs UV radiation. A dense cloud cover absorbs UV radiation, while a split cloud cover may tend to amplify it.</p><p>Although the stratospheric ozone layer is showing signs of recovery from reduced ozone-depleting substances. The impact of greenhouse gases on the climate is still in increase and global climate models anticipate an acceleration in Brewer-Dobson Circulation, which would lead to lower ozone levels in the tropics. Butler et al. (2016) estimate a decrease in stratospheric ozone in the tropics of 5 to 10 DU for all climate scenarios. Some recent projections (Lamy et al., 2019) predict a 2-3% increase in UVR in the southern tropical band, a region where UV levels are already extreme.</p><p>The purpose of the UV-Indien network is to :</p><p>- Monitor UV levels at different sites in the Western Indian Ocean (WIO)</p><p>- Describe the annual and inter-annual variability of UV radiation in the WIO</p><p>- Perform regional climate projections of UV radiations, validated by quality ground measurements.</p><p>UV-Indien is split into three phases. The first phase began in 2016, with the deployment of the first measurement sites (Reunion Island, Madagascar, Seychelles, Rodrigues). These sites are equipped with a broadband radiometer measuring the UVI and a camera estimating the coverage and sometimes a spectrometer for the measurement of total ozone. The second phase from 2019, sees the extension of this network to 4 other sites (Juan de Nova, Diego Suarez, Fort Dauphin and Grande Comoros). The data validation phase began in 2019 (comparative study with satellite data) and will also propose the study of the variability of UV radiation on different sites. Finally, climate projections will be made from 2020 onwards and will use data from the network to validate the results.</p><p>The aim of this communication is to describe the entire network and its objectives. The first results, as well as the first climatologies will also be discussed.</p>

<strong class="journal-contentHeaderColor">Abstract.</strong> Within the framework of the UV-Indien network, nine ground stations have been equipped with ultraviolet broadband radiometers, five of them have also been equipped with an all-sky camera, and the main station in Saint-Denis de la Réunion is also equipped with a spectroradiometer. These stations are spatially distributed to cover a wide range of latitudes, longitudes, altitudes, and environmental conditions in five countries of the western Indian Ocean region (Comoros, France, Madagascar, Mauritius, and Seychelles), a part of the world where almost no measurements have been made so far. The distribution of the stations is based on the scientific interest of studying ultraviolet radiation not only in relation to atmospheric processes but also in order to provide data relevant to fields such as biology, health (prevention of skin cancer), and agriculture. The main scientific objectives of this network are to study the annual and inter-annual variability in the ultraviolet (UV) radiation in this area, to validate the output of numerical models and satellite estimates of ground-based UV measurements, and to monitor UV radiation in the context of climate change and projected ozone depletion in this region. A calibration procedure including three types of calibrations responding to the various constraints of sustaining the network has been put in place, and a data processing chain has been set up to control the quality and the format of the files sent to the various data centres. A method of clear-sky filtering of the data is also applied. Here, we present an intercomparison with other datasets, as well as several daily or monthly representations of the UV index (UVI) and cloud fraction data, to discuss the quality of the data and their range of values for the older stations (Antananarivo, Anse Quitor, Mahé, and Saint-Denis). Ground-based measurements of the UVI are used to validate satellite estimates – Ozone Monitoring Instrument (OMI), the TROPOspheric Monitoring Instrument (TROPOMI), and the Global Ozone Monitoring Experiment (GOME) – and model forecasts of UVI – Tropospheric Emission Monitoring Internet Service (TEMIS) and Copernicus Atmospheric Monitoring Service (CAMS). The median relative differences between satellite or model estimates and ground-based measurements of clear-sky UVI range between <span class="inline-formula">−</span>34.5 % and 15.8 %. Under clear skies, the smallest UVI median difference between the satellite or model estimates and the measurements made by ground-based instruments is found to be 0.02 (TROPOMI), 0.04 (OMI), <span class="inline-formula">−</span>0.1 (CAMS), and <span class="inline-formula">−</span>0.4 (CAMS) at Saint-Denis, Antananarivo,<span id="page4276"/> Anse Quitor, and Mahé, respectively. The diurnal variability in UVI and cloud fraction, as well as the monthly variability in UVI, is evaluated to ensure the quality of the dataset. The data used in this study are available at <a href="https://doi.org/10.5281/zenodo.4811488">https://doi.org/10.5281/zenodo.4811488</a> <span class="cit" id="xref_paren.1">(<a href="#bib1.bibx28">Lamy and Portafaix</a>, <a href="#bib1.bibx28">2021</a><a href="#bib1.bibx28">a</a>)</span>.

<strong class="journal-contentHeaderColor">Abstract.</strong> Within the framework of the UV-Indien network, nine ground stations have been equipped with ultraviolet broadband radiometers, five of them have also been equipped with an all-sky camera, and the main station in Saint-Denis de la Réunion is also equipped with a spectroradiometer. These stations are spatially distributed to cover a wide range of latitudes, longitudes, altitudes, and environmental conditions in five countries of the western Indian Ocean region (Comoros, France, Madagascar, Mauritius, and Seychelles), a part of the world where almost no measurements have been made so far. The distribution of the stations is based on the scientific interest of studying ultraviolet radiation not only in relation to atmospheric processes but also in order to provide data relevant to fields such as biology, health (prevention of skin cancer), and agriculture. The main scientific objectives of this network are to study the annual and inter-annual variability in the ultraviolet (UV) radiation in this area, to validate the output of numerical models and satellite estimates of ground-based UV measurements, and to monitor UV radiation in the context of climate change and projected ozone depletion in this region. A calibration procedure including three types of calibrations responding to the various constraints of sustaining the network has been put in place, and a data processing chain has been set up to control the quality and the format of the files sent to the various data centres. A method of clear-sky filtering of the data is also applied. Here, we present an intercomparison with other datasets, as well as several daily or monthly representations of the UV index (UVI) and cloud fraction data, to discuss the quality of the data and their range of values for the older stations (Antananarivo, Anse Quitor, Mahé, and Saint-Denis). Ground-based measurements of the UVI are used to validate satellite estimates – Ozone Monitoring Instrument (OMI), the TROPOspheric Monitoring Instrument (TROPOMI), and the Global Ozone Monitoring Experiment (GOME) – and model forecasts of UVI – Tropospheric Emission Monitoring Internet Service (TEMIS) and Copernicus Atmospheric Monitoring Service (CAMS). The median relative differences between satellite or model estimates and ground-based measurements of clear-sky UVI range between <span class="inline-formula">−</span>34.5 % and 15.8 %. Under clear skies, the smallest UVI median difference between the satellite or model estimates and the measurements made by ground-based instruments is found to be 0.02 (TROPOMI), 0.04 (OMI), <span class="inline-formula">−</span>0.1 (CAMS), and <span class="inline-formula">−</span>0.4 (CAMS) at Saint-Denis, Antananarivo,<span id="page4276"/> Anse Quitor, and Mahé, respectively. The diurnal variability in UVI and cloud fraction, as well as the monthly variability in UVI, is evaluated to ensure the quality of the dataset. The data used in this study are available at <a href="https://doi.org/10.5281/zenodo.4811488">https://doi.org/10.5281/zenodo.4811488</a> <span class="cit" id="xref_paren.1">(<a href="#bib1.bibx28">Lamy and Portafaix</a>, <a href="#bib1.bibx28">2021</a><a href="#bib1.bibx28">a</a>)</span>.

The rapid degradation of ecosystems, exacerbated by human activities and climate change, threatens the provision of essential ecosystem services on which we depend. In light of this situation, this thesis aims to contribute to improved environmental management through the application of natural capital ecosystem accounting.This research applies ecosystem accounting of natural capital to the Mahavavy-Kinkony Complex, a protected area in northwestern Madagascar renowned for its ecological wealth but subject to intense anthropogenic pressures. Covering the period 2013–2018, the study combines satellite image analysis, the production of biophysical indicators, and the spatialization of results at the watershed scale. Five biophysical accounts were developed: land cover, carbon, water, ecosystem infrastructure, and capability.The results highlight significant landscape transformation, driven by agricultural expansion, in particular the growth of rice cultivation (+60%) and sugarcane (+26%), accompanied by a decline in carbon stocks, a reduction in water resources (lake storage decreased by 20%), and increasing fragmentation of natural habitats. Dry forests appear particularly vulnerable, whereas wetlands and mangroves show greater resilience to pressures. Overall, ecosystem capability declined across the entire area, with an ecological debt estimated at 271,506 ECU. The spatialization of accounts provides an integrated reading of territorial dynamics, supporting the identification of critical areas and the prioritization of conservation, restoration, and sustainable management actions.By demonstrating both the technical feasibility and the strategic relevance of applying ecosystem accounting of natural capital at the local scale, this thesis bridges global frameworks and territorial realities. Its findings provide an operational tool capable of informing the decisions of managers and public stakeholders concerning the preservation of natural capital.

A unique dataset of -Manning consistent- stage/discharge rating curves (RCs) was built at virtual stations (VSs) from the 27-day repeat cycle Sentinel3-A and 10-day repeat cycle Jason-3 missions. For this purpose, we used discharges simulated with the MGB hydrologic-hydrodynamic model and the satellite altimetry time series in several tropical and sub-tropical hydrological watersheds, including the Amazon and Maroni basins in South America and the Niger, Congo, Ogooue and Tsiribihina basins in Africa. The resulting RCs were used to get a priori information on the RC’s parameters at the newly released Sentinel3-B VSs. Thus, we get a dense framework of RCs, ranging from less than ten (for the Tsiribihina basin) to hundreds (for the Congo and Amazon basins) of VSs per basin. The RCs validation was threefold: 1) a systematic validation of the rated discharge against simulated discharge in overlapping time period; 2) a validation against in-situ discharge was performed when available; and 3) field works provided some precious comparison between estimated depths and/or rating curve parameter and ground truth. Whenever possible and physically acceptable, time series in the close vicinity of each other were merged in order to increase the frequency of revisit. Thanks to the operational status of the COPERNICUS Sentinel3-A&amp;B and NASA/CNES Jason3 satellites, users are ensured to benefit from new observations of water height, hence from new estimates of depth and discharge, during the all mission lifetime, i.e. at least 15 years. Discharge information will shortly be released on the Hydroweb website (http://hydroweb.theia-land.fr/) under the form of the RC coefficients to be applied to the water surface elevation time series.