Nowadays, modern earth observation programs produce huge volumes of satellite images time series (SITS) that can be useful to monitor geographical areas through time. How to efficiently analyze such kind of information is still an open question in the remote sensing field. Recently, deep learning methods proved suitable to deal with remote sensing data mainly for scene classification (i.e. Convolutional Neural Networks -CNNs -on single images) while only very few studies exist involving temporal deep learning approaches (i.e Recurrent Neural Networks -RNNs) to deal with remote sensing time series.In this letter we evaluate the ability of Recurrent Neural Networks, in particular the Long-Short Term Memory (LSTM) model, to perform land cover classification considering multi-temporal spatial data derived from a time series of satellite images. We carried out experiments on two different datasets considering both pixel-based and object-based classification. The obtained results show that Recurrent Neural Networks are competitive compared to state-of-the-art classifiers, and may outperform classical approaches in presence of low represented and/or highly mixed classes. We also show that using the alternative feature representation generated by LSTM can improve the performances of standard classifiers.

Purpose Since the implementation of the European directive (EC/2001/42) on strategic environmental assessment, an ex ante evaluation has become mandatory for plans and programs. This requirement could have significant consequences for the environment. Local authorities, who are in charge of land planning issues, must therefore conduct such assessments. However, they are faced with lack of uniform methodology. The aim of this paper is thus to propose a methodological framework for the required environmental assessment stages in land planning. Methods Life cycle assessment (LCA) has been identified as a promising tool to perform environmental assessment at a meso-level (i.e., territories). Yet, the standardized LCA framework has never been used for assessing the environmental impacts of a territory as such, which can be explained by the complexity that its application would involve. Four major methodological bottlenecks have been identified in this paper, i.e., (1) functional unit definition, (2) boundary selection, (3) data collecting, and (4) the refinement of the life cycle impact assessment phase in order to provide useful indicators for land planning. For each of these challenges, recommendations have been made to adapt the analytical framework of LCA. Results and discussion A revised framework is proposed to perform LCA of a territory. One of the major adaptations needed concerns the goal and scope definition phase. Henceforth, the association of a territory and the studied land planning scenario, defined by its geographical boundaries and its interactions with other territories, will be designated as the reference flow in LCA. Consequently, two kinds of indicators will be determined using this approach, i.e., (1) a vector of environmental impacts generated (conventional LCA) and (2) a vector of land use functions provided by the territory for different stakeholders (e.g., provision of work, recreation, culture, etc.). This revised framework has been applied to a theoretical case study in order to highlight its utility in land planning. Conclusions This work is a first step in the adaptation of the LCA framework to environmental assessment in land planning. We believe that this revised framework has the potential to provide relevant information in decision-making processes. Nonetheless, further work is still needed to broaden and deepen this approach (i.e., normalization of impacts and functions, coupled application with GIS, uncertainties, etc.).

The partitioning of aluminium between silicate melt (olivine tholeiite) and chromian spinel was studied experimentally. Runs were carried out at temperatures of 1 180˗1 300 °C and oxygen fugacities of 10⁻⁷˗10⁻⁹ atmosphere, at one atmosphere total pressure. The Al₂O₃ content of spinel is related to the Al₂O₃ content of melt as follows : (Al₂O₃)[sp] = 0.035 (Al₂O₃)[liq], where Al₂O₃ is expressed in weight %. Thus there is a positive correlation between the partition coefficient D = (Al₂O₃)[sp]/(Al₂O₃)[liq] and the Al₂O₃ content of the melt. Temperature and oxygen fugacity have only a minor effect on D. Consequently, within the limits of experimental conditions, the chromium content of spinel increases as the Al₂O₃ content of the melt decreases.

No abstract

No abstract