Land cover and land use maps derived from satellite remote sensing imagery are critical to support biodiversity and conservation, especially over large areas. With its 10 m to 20 m spatial resolution, Sentinel-2 is a promising sensor for the detection of a variety of landscape features of ecological relevance. However, many components of the ecological network are still smaller than the 10 m pixel, i.e., they are sub-pixel targets that stretch the sensor’s resolution to its limit. This paper proposes a framework to empirically estimate the minimum object size for an accurate detection of a set of structuring landscape foreground/background pairs. The developed method combines a spectral separability analysis and an empirical point spread function estimation for Sentinel-2. The same approach was also applied to Landsat-8 and SPOT-5 (Take 5), which can be considered as similar in terms of spectral definition and spatial resolution, respectively. Results show that Sentinel-2 performs consistently on both aspects. A large number of indices have been tested along with the individual spectral bands and target discrimination was possible in all but one case. Overall, results for Sentinel-2 highlight the critical importance of a good compromise between the spatial and spectral resolution. For instance, the Sentinel-2 roads detection limit was of 3 m and small water bodies are separable with a diameter larger than 11 m. In addition, the analysis of spectral mixtures draws attention to the uneven sensitivity of a variety of spectral indices. The proposed framework could be implemented to assess the fitness for purpose of future sensors within a large range of applications.

Soil fertility in smallholder farming areas is known to vary strongly on multiple scales. This study measures the sensitivity of the recorded satellite signal to on-farm soil fertility treatments applied to five crop types, and quantifies this fertilization effect with respect to within-field variation, between-field variation and field position in the catena. Plant growth was assessed in 5-6 plots per field in 48 fields located in the Sudano-Sahelian agro-ecological zone of southeastern Mali. A unique series of Very High Resolution (VHR) satellite and Unmanned Aerial Vehicle (UAV) images were used to calculate the Normalized Difference Vegetation Index (NDVI). In this experiment, for half of the fields at least 50% of the NDVI variance within a field was due to fertilization. Moreover, the sensitivity of NDVI to fertilizer application was crop-dependent and varied through the season, with optima at the end of August for peanut and cotton and early October for sorghum and maize. The influence of fertilizer on NDVI was comparatively small at the landscape scale (up to 35% of total variation), relative to the influence of other components of variation such as field management and catena position. The NDVI response could only partially be benchmarked against a fertilization reference within the field. We conclude that comparisons of the spatial and temporal responses of NDVI, with respect to fertilization and crop management, requires a stratification of soil catena-related crop growth conditions at the landscape scale.

Changes in the snow cover extent are both a cause and a consequence of climate change. Optical remote sensing with heliosynchronous satellites currently provides snow cover data at high spatial resolution with daily revisiting time. However, high latitude image acquisition is limited because reflective sensors of many satellites are switched off at high solar zenith angles (SZA) due to lower signal quality. In this study, the relevance and reliability of high SZA acquisition are objectively quantified in the purpose of high latitude snow cover detection, thanks to the PROBA-V (Project for On-Board Autonomy-Vegetation) satellite. A snow cover extent classification based on Normalized Difference Snow Index (NDSI) and Normalized Difference Vegetation Index (NDVI) has been performed for the northern hemisphere on latitudes between 55 • N and 75 • N during the 2015-2016 winter season. A stratified probabilistic sampling was used to estimate the classification accuracy. The latter has been evaluated among eight SZA intervals to determine the maximum usable angle. The global overall snow classification accuracy with PROBA-V, 82% ± 4%, was significantly larger than the MODIS (Moderate-resolution Imaging Spectroradiometer) snow cover extent product (75% ± 4%). User and producer accuracy of snow are above standards and overall accuracy is stable until 88.5 • SZA. These results demonstrate that optical remote sensing data can still be used with large SZA. Considering the relevance of snow cover mapping for ecology and climatology, the data acquisition at high solar zenith angles should be continued by PROBA-V.