 |
| Titel |
Data mining of external and internal forcing of fluvial systems for catchment management: A case study on the Red River (Song Hong), Vietnam |
| VerfasserIn |
Rafael Schmitt, Simone Bizzi, Andrea Castelletti |
| Konferenz |
EGU General Assembly 2013
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 15 (2013) |
| Datensatznummer |
250082705
|
|
|
|
|
|
| Zusammenfassung |
| The understanding of river hydromorphological processes has been recognized in the last
decades as a priority of modern catchment management, since interactions of natural and
anthropogenic forces within the catchment drives fluvial geomorphic processes, which shape
physical habitat, affect river infrastructures and influence freshwater ecological processes.
The characterization of river hydromorphological features is commonly location and time
specific and highly resource demanding. Therefore, its routine application at regional or
national scales and the assessment of spatio-temporal changes as reaction to internal and
external disturbances is rarely feasible at present. Information ranging from recently available
high-resolution remote-sensing data (such as DEM), historic data such as land use maps or
aerial photographs and monitoring networks of flow and rainfall, open up novel and
promising capacity for basin-wide understanding of dominant hydromorphological
drivers. Analysing the resulting multiparametric data sets in their temporal and spatial
dimensions requires sophisticated data mining tools to exploit the potential of this
information. We propose a novel framework that allows for the quantitative assessment of
multiparametric data sets to identify classes of channel reaches characterized by
similar geomorphic drivers using remote-sensing data and monitoring networks
available in the catchment. This generic framework was applied to the Red River
(Song Hong) basin, the second largest basin (87,800 sq.km) in Vietnam. Besides its
economic importance, the river is experiencing severe river bed incisions due to recent
construction of new dams in the upstream part of the catchment and sand mining
in the surrounding of the capital city Hanoi. In this context, characterized by an
high development rate, current efforts to increase water productivity and minimize
impacts on the fluvial systems by means of focused infrastructure and management
measures require a thorough understanding of the fluvial system and, in particular,
basin-wide assessment of resilience to human-induced change. . The framework
proposed has allowed producing high-dimensional samples of spatially distributed
geomorphic drivers at catchment scale while integrating recent and historic point
records for the Red River basin. This novel dataset has been then analysed using
self-organizing maps (SOM) an artificial neural network model in combination with
fuzzy clustering. The above framework is able to identify non-trivial correlations in
driving forces and to derive a fuzzy classification at reach scale which represents
continuities and discontinuities in the river systems. The use of the above framework
allowed analyzing the spatial distribution of geomorphic features at catchment scale,
revealing patterns of similarities and dissimilarities within the catchment and allowing a
classification of river reaches characterized by similar geomorphic drivers, fluvial
processes and response to external forcing. The paper proposes an innovative and
promising technique to produce hydromorphological classifications at catchment scale
integrating historical and recent available high resolution data. The framework
aims at opening the way to a more structured organization and analyses of recently
available information on river geomorphic features, so far often missing or rarely
exploited. This approach poses the basis to produce efficient databases of river
geomorphic features and processes related to natural and anthropogenic drivers. That
is a necessity in order to enhance our understanding of the internal and external
forces which drive fluvial systems, to assess the resilience and dynamic of river
landscapes and to develop the more efficient river management strategies of the
future. |
|
|
|
|
|
|