Science Service System

Summary of Proposal LAN0922

TitleSynergistic use of Radarsat-2 and TerrSAR-X for monitoring vegetation distribution, growth and phenology in the Amazon floodplain wetlands
Investigator Silva, Thiago - São Paulo State University (UNESP), Geography Department, Rio Claro campus
Team Member
Dr. Costa, Maycira - University of Victoria, Geography
Dr. Melack, John - University of California Santa Barbara, Bren School of Environmental Science and Management
Dr. Novo, Evlyn - National Institute for Space Research, Remote Sensing Division
Mrs. Sartori, Lauriana - São Paulo State University (UNESP), Graduate Program in Cartographic Sciences
SummaryThe ecological dynamics of the Amazon wetlands are determined by the annual flooding cycle. Studies predict an increase in the frequency and intensity of droughts for the region due to climate change, which will in turn affect vegetation dynamics and their role in floodplain ecology and biogeochemistry. The co-existence of full polarimetric L-band, C-Band and X-band orbital SAR systems offers unprecedented possibilities for monitoring vegetation dynamics in the Amazon floodplain. The present study thus proposes to evaluate the synergistic use of ALOS Palsar, Radarsat 2 and TerraSAR-X to study vegetation processes in the Amazon floodplain, through the following objectives: 1) Establish an operational method, based on combined L, C and X band data, map and monitor vegetation cover and phenology; 2) Evaluate the capacity of full polarimetric data to discriminate plant species and estimate plant biophysical parameters 3) Evaluate the effectiveness of ScanSAR data for large scale vegetation mapping, compared to standard resolution products; 4) Evaluate the suitability of high-resolution X band images for visual interpretation and ground-truthing for inaccessible areas of the floodplain. This research will rely on the coordinated acquisition of SAR data, complemented by two to four missions to collect ground data regarding vegetation types, main species, plant biomass and distribution. The first step in the study will consist in the analysis of the backscattering temporal signature for all vegetation types, supported by ground truth data. This will allow a complete understanding of the effects of wavelength, polarization and spatial resolution on the signatures, and the relationships between backscattering and phenological changes. The second step will be then to evaluate the effectiveness of combined FBD ALOS Palsar images, Standard Beam Radarsat-2 and ScanSAR TerraSAR-X images to map floodplain vegetation communities over time. In this step, the method originally proposed by Silva et. al. (2010) will be adapted to incorporate the three image sources. Also during this step, the usefulness of high resolution TerraSAR-X imagery for visual interpretation and validation will be assessed. The third step will evaluate of the trade-offs between spatial resolution and image coverage, by utilizing the same classification algorithm to classify areas covered by ALOS Palsar and Radarsat-2 ScanSAR scenes. Finally,a detailed analysis of full polarimetric ALOS Palsar, Radarsat-2 and TerraSAR-X products will be conducted. A range of incoherent and coherent attributes will be calculated and associated with field data, to quantify the capacity to discriminate plant species and accurately estimate biophysical vegetation attributes. The proposed study area is located within 54º00’W -56º00’W and 01º50’S - 02º30’S, Pará state, Brazil. This area have a long track-record of previous studies and recorded data on hydrology, water chemistry, and plant dynamics. For the proposed research, the following TerraSAR-X image acquisitions are requested: ScanSAR: Up to 12 scenes, acquired in near-monthly intervals, covering an entire seasonal cycle (December 2010 - December 2011), within scheduling constraints. ScanSAR TerraSAR-X images will complement Standard Mode Radarsat-2images for monitoring vegetation cover. Dual-polarization, Stripmap: 2 scenes,timed with the extremes in plant phenology, corresponding to high (May - July)and low (October - December) water periods in the seasonal cycle. Spotlight SL mode: 2 scenes, timed with the extremes in plant phenology, as above. The proposed research will be supported by three currently funded grants: NSERC 243099 (P.I.: Maycira Costa); NASA NNX10AB66G (P.I.: John Melack); CNPq 550373/2010-1 and FAPESP 2008/07537-1 (P.I.: Evlyn Novo). These grants will provide financial, logistical and technological support for the acquisition and analysis of field and image data.

Back to list of proposals

© DLR 2004-2016