Autori: Ruxandra Vintila, Catalin Lazar, Violeta Poenaru, Cristina Radnea, Petre Voicu
Editorial: ISSN: 0585-3052, Romanian National Society of Soil Science, Stiinta Solului – Soil Science (Journal of the Romanian National Society of Soil Science), XLVI (1), p.33-52, 2012.
The present article constitutes a review of the innovative approaches and results obtained in the ADAM project, dedicated to the Assimilation of spatial Data into Agronomic Models, a scientific collaboration between France and Romania. The results mainly refer to the following aspects: (i) constitution of the first reference spatial remote sensing and agronomic knowledge base for scientific investigation; (ii) production of a SPOT XS/XI time series of high quality satellite images; (iii) validation of a method for monitoring soil surface moisture throughout crop phenological cycles, using SAR (ERS-2 and RADARSAT-1) images and the water cloud model; (iv) definition of the revisit frequency of satellites for the field-scale agriculture; (v) development of an efficient strategy of variational assimilation of spatial data into agronomic models, by exploiting the high spatial coherence that characterizes the crops during their development; (vi) calculation of the adjoint model of the complex canopy functioning model STICS by automatic differentiation; (vii) improvement of the canopy radiative transfer modeling by accounting for the leaf clumping, and elaboration of the CLAMP model. Furthermore, this article presents other studies, led to valorize
the ADAM knowledge after this project had finished. These are primarily related to the development of pattern analysis algorithms (i.e., advanced data mining for efficient extraction of information on spatio-temporal phenomena from Satellite Image Time Series / SITS, and data fusion for multi-resolution decomposition, by using the morphological pyramid technique), which the high quality of the ADAM knowledge base made possible.
Being free for scientific studies, with easy access through the Kalideos Portal, the ADAM knowledge base still has the potential to produce other notable findings (http://kalideos.cnes.fr)./spip.php?article68).