This study analyses a long-time series (from January 2016 to December 2021) of optical and SAR signatures. The data were processed into two contexts: 1) over a station-monitored field called Auradé in southwestern France (part of the ICOS network) to examine the impact of surface states on satellite signals and, 2) over 939 fields of sunflower to investigate the impact of climatic conditions on crop development. Results show that backscatters can be used with moderate confidence to replace or gap fill NDVI ($r_{6-years}^{Aurad\acute{e}} = 0.73$ over the Auradé field and $r_{6 - years}^{sunflower} = 0.77$ over the 939 fields of sunflower). Nevertheless, additional analyses highlighted the effectiveness of combining γ0VH, γ0VH/VV with NDVI, GAI (Green Area Index), and Fcover indices for complementary monitoring of crop phenology. Superpositions of satellite signals observed on sunflowers suggest that it is possible to detect development anomalies using optical or radar signals based on yearly climatic conditions. Results also show the importance of considering both sunflower architecture (orientation, inter-row spacing, Fcover) and radar acquisition geometry (orbit pass, viewing angles), in contrast to denser or more covering crops. The joint use of the 2 SAR orbits enabled the detection of phenological stages (from flowering to harvest) not as well detectable with a single orbit. This long-time series analysis provided insights under multiple climatic conditions, including the hottest year in France since the beginning of meteorological records.

Since 2016, land surfaces can be monitored from optical and synthetic aperture radar sensors onboard Sentinel-1 and Sentinel-2 satellites at high spatial and temporal resolutions. Monitoring agricultural surfaces through satellite-based estimations of biophysical parameters is a key issue for agriculture sustainability in the context of an increasing climate change. With its all-weather vision capability, it is important to develop methods based on the use of Sentinel-1 images. In this study, we applied an original algorithm of change detection to monitor the two summer crops (sunflower and sorghum) grown in France and other parts according to a near-real-time detection method called CuSum-NRT applied to Sentinel-1 time-series of images.

We present a practical approach to inter-compare a range of candidate digital elevation models (DEMs) based on pre-defined criteria and statistically sound ranking approach. The presented approach integrates the randomized complete block design (RCBD) into a novel framework for DEMs comparison. The method presented provides a flexible, statistically sound and customizable tool for evaluating the quality of any raster - in this case a DEM - by means of a ranking approach, which takes into account a confidence level, and can use both quantitative and qualitative criteria. The users can design their own criteria for the quality evaluation in relation to their specific needs. The application of the RCBD method to rank six 1” global DEMs, considering a wide set of study sites, covering different morphological and landcover settings, highlights the potentialities of the approach. We used a suite of criteria relating to the differences in the elevation, slope, and roughness distributions compared to reference DEMs aggregated from 1-5 m lidar-derived DEMs. Results confirmed significant superiority of CopDEM 1” and its derivative FABDEM as the overall best 1” global DEMs. They are slightly better than ALOS, and clearly outperform NASADEM and SRTM, which are in turn much better than ASTER.

Tropical forests are vulnerable to deforestation. This phenomenon led to the development of a wide number of forest monitoring systems based on remotely sensed data. These systems face multiple issues in tropical areas, one of them being the need of a good forest / non-forest map to use as reference for the monitoring. Several of these maps are available online, most of them being based on optical remote sensing data, which is known to be subject to limitations in tropical areas due to high cloud cover. Sentinel-1 C-band Synthetic Aperture Radar (SAR) dense time series of images has already been used for forest / non-forest mapping through combination with X-band SAR images. In this study, a change detection algorithm was applied on timeseries of Sentinel-1 images in the Parà State, Brazil. The Cumulative Sum (CuSum) algorithm was used to assess all non-forest areas as the hypothesis found by preliminary results was that these areas were more varying in terms of backscatter coefficient than undisturbed forests. The validation was made by comparison with a forest / non-forest map derived from Global Forest Watch data. The algorithm detected 78.0 % of the non-forested areas in the study zone of 6,855 km² with a 0.785 F1-score value.

Lakes and reservoirs are considered sentinels of climate and anthropogenic changes. Lakes and reservoirs surface water storage is an essential hydrological variable but poorly known as this information is scarce. Earth Observation data are a reliable source of information to overcome this scarcity. Among these, the combined use of satellite images, to derive water extent, and radar altimetry, which enables to estimate water levels, provides valuable information on water storage changes. Here, we used Synthetic Aperture Radar images from Sentinel-1 and radar altimetry data from Sentinel-3 to monitor the water volume changes of Lake Bracciano from 2016 to 2021. This lake was affected by a water crisis in 2017 and the water supply to the city of Rome (Italy) was interrupted September 2017 to preserve its ecosystem. Hence, we demonstrate how Sentinel-1 and Sentinel-3 data can be useful to monitor water extent and level, which can be profoundly changed by the climate crisis.

Earth Observation data is often used for land cover classification or change monitoring. It is rarely used for both goals in a single algorithm. The multi-change Cumulative Sum (CuSum) algorithm proposed in this study allows both classification and change monitoring in a single algorithm using Sentinel-1 C-SAR time series. The multi-change CuSum approach allowed to classify pixels belonging to the fused non-forest vegetation and bare soil classes apart from the pixels belonging to new cuts. The distinction of each class is better made using the two polarizations: VV is more accurate for detecting non-forest vegetation (Kappa coefficient of 0.62) and VH for detecting new cuts (Kappa coefficient of 0.65). The algorithm showed an accuracy up to 0.82.

Tropical forests are vulnerable to deforestation and various monitoring techniques have been developed based on remotely sensed data to map deforestation, but are facing multiple problems in the tropical areas. For instance, the techniques based optical data, which are widely used to monitor deforestation, face severe limitations in the humid tropical forest due to high cloud cover. Sentinel-l C-SAR dense time series can be used for a temporally more accurate monitoring. In this study, a change detection algorithm commonly used in the financial domain, the Cumulative Sum (CuSum) algorithm, was modified to be applied on time-series of Sentinel-l images in a forest concession of Democratic Republic of Congo (DRC) near Kisangani. The validation was made through the visual interpretation of PlanetScope OrthoScene images as in-situ data were missing. The results show a precision up to 0.75, an accuracy up to 0.95 and a kappa coefficient up to 0.40 for clear cut detection. The algorithm is able to detect forest degradation activities before the clear cuts.

The aim of this paper was to propose a new methodology for preprocessing and enhancing C-band synthetic aperture radar (SAR) images for the automatic detection of marine oil slicks. The proposed methodology includes three processing levels: preprocessing, thresholding, and binary cleaning. The first level is to correct the heterogeneity of brightness in SAR images caused by the non-Lambertian reflection of the radar signal on the sea surface. This heterogeneity can be justified by: the distance from the nadir (incidence angle effect), the interaction between wind direction and radar pulse, and the wide swath mode. The second level consists of a thresholding step. The third level is to clean the binary output images from noise residues. Several preprocessing and cleaning methods have been tested and evaluated by a qualification engine that compares the automatically detected patches with a training data set of manually detected dark patches. The training data set includes oil slicks and lookalikes. As a result, the “best” preprocessing method that homogenizes the brightness of C-band SAR scenes and optimizes the automatic detection of marine oil slicks is based on an adaptation to the C-band MODel. As for the cleaning process, the tested morphological methods show that small object removal followed by a morphological closing optimizes the automatic detection of marine oil slicks.