Capabilities of BIOMASS Tomography for Investigating Tropical Forests
Ho Tong Minh, Dinh1; Tebaldini, Stefano2; Rocca, Fabio2; Le toan, Thuy1
1CESBIO, FRANCE; 2Politecnico di Milano, ITALY

The candidate Earth Explorer mission BIOMASS was presented at the User Consultation Meeting in March 2013 and recommended by the ESA's Earth Science Advisory Committee (ESAC) concerning further developments and eventual deployment. The satellite will be operated in two different observation phases, a tomographic phase, where multiple baselines will be collected with a revisit time of few days, and a nominal interferometric phase [1]. SAR Tomography (TomoSAR) imaging will allow to retrieve the vertical structure of the vegetation, which would be one of the key elements for the assessment of the forest biomass [2]. However, a major contributor to the error budget is the bandwidth limitation imposed to BIOMASS by ITU regulations, which allow for no more than 6 MHz [3]. Such a bandwidth limitation causes the SAR resolution cell to significantly spread along the Line of Sight (LOS). At the proposed incidence angle of 23° this translates into an appreciable resolution loss not only in the ground range direction, but in the vertical direction a well [4]. As a result, BIOMASS tomography is hindered by two different factors compared to airborne tomography, that is: i) A significant reduction of the number of looks to be used for coherence evaluation ii) A significant vertical resolution loss. The objective of this paper is to provide a better understanding of BIOMASS capabilities concerning the estimation of forest biomass and height by means of tomographic techniques. The analysis presented in this paper is carried out on airborne data acquired by ONERA over the site of Paracou, French Guyana, during the ESA campaign TropiSAR [5]. Those data have been reprocessed in order to generate a new data stack consistent with BIOMASS as for bandwidth and carrier frequency. To do this, two different processing approaches have been considered. One consisted in degrading the resolution of airborne data through linear filtering. This approach has the main advantage of being fast, although it does not allow to have the same LOS as the emulated spaceborne system along the whole imaged swath. The other approach consisted in recovering the 3D distribution of the scatterers at high resolution, which was then reprojected onto BIOMASS geometry accounting for the available radiofrequency bandwidth. This procedure allows to obtain a data stack that is consistent with BIOMASS concerning not only spatial resolution, but also geometrical features, i.e.: system LOS. Accordingly, the data stack obtained by reprojection exhibits the same vertical resolution along the swath, resulting in a faithful emulation of BIOMASS imaging capabilities. The connection to forest biomass has been examined in both cases, by investigating the correlation between backscattered power at different forest heights and above ground biomass (ABG) values from in-situ data. As expected, the reduction of system bandwidth to 6 MHz resulted in significant vertical resolution losses compared to the original airborne data (125 MHz). Nevertheless, it was possible to retrieve forest height to within an accuracy of less than 4 m, whereas the backscattered power at volume height (say 30 m above the ground) exhibited a correlation higher than 0.7 with in-situ data and no bias phenomena over ABG values ranging from 250 t/ha to 450 t/ha. The analysis was also extended in order to include temporal decorrelation effects. Temporal decorrelation parameters were extracted using data from the TropiScatt campaign [6], where a ground-based system was employed to study short and long temporal decorrelation at the forest site of Paracou, that is the same one investigated during TropiSAR. The ground-based system was designed in such a way as to provide 2D (range - height) resolution capabilities through the coherent combination of the signal from different antennas via tomographic techniques. By comparing, again coherently, tomographic images taken at different times it was possible to gain access to the variation of temporal coherence with respect to forest height. The results from this study show that BIOMASS tomography in tropical forest is feasible with a revisit pass time on the order of about 4 days. The information about the forest vertical structure is partly preserved even with revisit time on the order of 30 days. In this case, however, models should be developed to try and compensate for temporal decorrelation in TomoSAR imaging.

References: [1] European Space Agency. (2012) Reports for Mission Selection: Biomass, ESA SP-1324(1), European Space Agency, Noordwijk, The Netherlands, ISSN 0379-6566, 193pp, available at [2] D. Ho Tong Minh, T. Le Toan, F. Rocca, S. Tebaldini, M. Mariotti d'Alessandro, and L. Villard, "Relating P-Band SAR tomography to tropical forest biomass," Geoscience and Remote Sensing, IEEE Transactions on, Submitted, 2012. [3] ITU. (2004) ITU-2004 Article 5 ("Frequency Allocations") of the Radio Regulations, Edition 2004, International Telecommunication Union. [4] Tebaldini, S.; Rocca, F.; , "Forest structure from longer wavelength SARS," Geoscience and Remote Sensing Symposium (IGARSS), 2010 IEEE International , vol., no., pp.158-161, 25-30 July 2010 [5] P. C. Dubois-Fernandez, T. Le Toan, S. Daniel, H. Oriot, J. Chave, L. Blanc,L. Villard, M. W. J. Davidson, and M. Petit, "The tropisar airborne campaign in french guiana: Objectives, description, and observed temporal behavior of the backscatter signal," Geoscience and Remote Sensing, IEEE Transactions on, vol. PP, no. 99, pp. 1 - 14, 2012. [6] D. Ho Tong Minh, S. Tebaldini, F. Rocca, T. Koleck, P. Borderies, C. Albinet,L. Villard, A. Hamadi, and T. Le Toan , "A Ground Based Array for Tomographic Imaging of a Tropical Forest at P-Band", TGRS, accepted for publicationS, accepted for publication