|TSM/TDM Science Team Meeting 2016|
|Science Team Meetings|
|How to Submit a Proposal|
|COFUR Price List (scientific use)|
|TanDEM-X Science Service System|
|Title||TerraSAR-X data for hydrological modelling in headwater river systems of alpine mountain massifs|
|Investigator||Flügel, Wolfgang-Albert - Friedrich-Schiller-University of Jena (FSU-Jena), Geoinformatics, Hydrology and Modelling|
The overall objective of the proposal is to upgrade knowledge of IWRM in headwater river systems of alpine mountain massifs by integrating RS techniques. Professional expertise, approaches and tools relevant to IWRM shall be transferred from a European river basin into the twinning Asian basin enhancing institutional capacity and creating regional awareness.
The methodical approach comprises comprehensive hydrological and environmental systems analyses and assessment of the natural environment comprising e.g. runoff generation, glaciers and snow coverage, terrain, land use and land cover classification.
TS-X data will scientifically be evaluated in terms of the usability for (i) mapping and monitoring of glaciers and snow coverage, and (ii) a medium scale LULC classification. The project will result in a detailed description of how competitive the data will be compared to optical remotely sensed imagery from sensors like Landsat TM/ETM or Terra ASTER? Can TS-X data be used as a stand alone product for glacier and snow cover mapping or LULC classification? Can additional information be retrieved which cannot be derived from other data applied in the project? Those questions will be answered as an essential part of the project.
The information derived from those RS analyses will be used for modelling of the hydrological water balance for the Inn River Basin and adaption of the hydrological model to the monsoonal Upper Brahmaputra River Basin environment. This includes their snow melt driven hydrological dynamics to specify present and predict likely periods of floods and water stress.
The outcome of the project will exemplify a detailed estimation of the water regimes in river basins. The workflow based on TS-X remote sensing data will be a standardized product; the approach will need some tuning and then it should be directly transferable to other river catchments. At this point commercial companies could take benefit of the method and may use the algorithm for the development of ‘value added products’ all based on the TS-X data.
TS-X data requirements will be as follows:
Expected deliverables of the project are as follows:
Funding for this TS-X pre-launch AO proposal has not yet been allocated but application will be made in a project addressed to the EU: In Nov. 2005 a proposal for a specific targeted research project (STREP) will be submitted to the European Commission’s Sixth Framework Program (FP) in the thematic area “Sustainable Development, Global Change and Ecosystems”. The proposed TS-X pre-launch AO proposal covers a part of this EU proposal, and all partners of the TS-X proposal are also partners of the intended STREP. The EU project’s duration is proposed for 3 years and will start in case of acceptance mid of 2006.
|Detailed report||Introduction: The overall objective of this research is to upgrade knowledge of Integrated Water Resources Management in headwater river systems of alpine mountain massifs by integrating Earth Observation techniques. Case studies are carried out in the macro-scale basin of Upper Brahmaputra River in Southeast Asia which has extensive wetland areas in the alpine mountain headwaters and in the flood plains. These wetlands are unique in their biodiversity and strongly depend on the hydrological dynamics of the system. During monsoon no cloud free optical satellite data can be obtained. Therefore TerraSAR-X Stripmap and ScanSAR radar data are investigated for the assessment and monitoring of the wetland distribution. The scientific objective of this research is the evaluation of TerraSAR-X data in terms of usability for a medium scale Land Use Land Cover (LULC) classification. Increasing frequency and quality of remote sensing data created strong demand for high-quality retrieval of thematic information as input for analysis, modelling and decision-making purposes. Furthermore there is growing demand to integrate the processes associated with analysis of remote sensing imagery into workflows of landscape management systems. To meet those demands, research has been carried out to develop new methods for image analysis and classification at the Department of Geoinformatics, Hydrology and Modelling at FSU Jena based on IMALYS software solution developed by H.G. Geo Data Solutions GmbH company. Test Site Description: The test site covers a part of Brahmaputra River basin in Assam state, north-east India. This part of Brahmaputra River basin is characterized as a flood plain area with some hilly region in southern part. The Brahmaputra River, which flows through Assam for about 640 km, along with its tributaries produces a large number of floodplain wetlands. These wetlands are in the form of lakes, oxbow lakes/cut-off meanders, swamp/marshes, water logged area, reservoirs, and tanks. They represent a vast sheet of water with varying shape, size and depth. The wetlands of Assam state are home to a large variety of flora and fauna, some of which are extremely rare and endangered while others are of great ecological and economic value. Methodological Approach: TerraSAR-X Stripmap (TS-X SM) data in VV and HH polarization are investigated acquired on 25th February 2008, 8th May 2008 (beginning of monsoon in the study area), and 2nd July 2008. The ground resolution of the radar scenes is app. 6 meter. Object oriented land cover classification is carried out using IMALYS software which has been designed as a tool for object based image analysis using semantic and contextual information. IMALYS has been designed to also allow users who are not remote sensing experts to extract information from remotely sensed data according to the requirements of their individual tasks. Its main features are as follows: a. IMALYS implements various algorithms for the aggregation of image pixels into regions of homogeneous properties, i.e. segmentation. Within the scope of IMALYS these regions are referred to as ‘cells’. It is possible to derive cells without almost any user interaction by exclusively using image data statistics. b. IMALYS retrieves color, texture and shape characteristics for the derived cells. It further builds a topology, i.e. registers the spatial relationship between cells and stores both attributes and topology in a database table. c. IMALYS implements different classification methods that use the combination of both derived cells and their associated attributes to identify real-world objects in satellite imagery. The classification is optimized with respect to the number of used input features and feature combinations and is conducted according to provided reference data. During image segmentation, statistically homogeneous regions (cells) are delineated which are the basis for the following object-based analysis. The algorithm allows to control size and distribution of the cells by just a small number of parameters, allowing to adjust the cell size to the size of the expected land-cover elements. The feature retrieval step serves the calculation of integrated attributes for each cell based on its original image data (i.e. pixel values), its geometry (i.e. form and size), statistical and textural attributes calculated within the cells and optional additional data layers. All following classification steps are based on these calculated cell attributes. The final classification is subdivided into two steps. The first step combines cells with broadly similar image features by means of an unsupervised classification approach (clustering). The second step uses reference areas and supervised classification including spatial combinations of the previously generated clusters to produce the final object-based classification result. Results: A medium scale LULC classification was derived from TerraSAR-X Stripmap data using an object oriented classification approach. Consequently, TS-X Stripmap data can be used as a stand-alone product for a medium scale LULC classification even during monsoon season.|
Back to list of proposals
|© DLR 2004-2016||