TU Berlin

Geoinformation in Environmental PlanningMoKka - Modelling the distribution of organic carbon stocks in floodplain soils (DFG)

Page Content

to Navigation

MoKka - Modelling the distribution of organic carbon stocks in floodplain soils (DFG)

Modelling the distribution of organic carbon stocks in floodplain soils with VHR remote sensing data and additional geoinformation
German Research Foundation (DFG)
12/1/2009 - 11/30/2011
Research Associate:
Mag. rer. nat Leonhard Suchenwirth
Project Lead:
Prof. Dr. Birgit Kleinschmit


Floodplain soils play a crucial role when storing carbon; but there are few data available on the carbon stocks in these soils in comparison to other terrestrial ecosystems. Remote sensing data have been used for quite some time for the detection of soil characteristics. However, there is still no scientific basis for the generation of large scale soil maps showing the distribution of organic carbons in floodplain soils which are based on remote sensing and additional data.
The research area is the Donauauen National Park, which is situated along the river Danube between Vienna, Austria and Bratislava, Slovakia. It preserves the last remaining major wetlands environment in Central Europe. Here, the Danube is still free flowing and is the lifeline of the National Park.
In a first approach the variables water regime, the vegetation, the relief position and the content of clay and iron oxides were identified to have an effect on the carbon content. Two different types of sedimentation areas can be distinguished: a) Dynamic areas, stocked with younger trees and a higher number of trees, and due to frequent inundations with a high flow velocity a higher number of soil horizons; and b) stable areas with a lower number of trees, a higher distance to the river, and due to a lower flow velocity during inundations a lower number of soil horizons. On aerial images, five vegetation units could be differentiated visually: Salix alba, softwood floodplain forests, populus alba, populus x canadensis and hardwood floodplain forests.
Based on these findings the overall goal is to develop a knowledge-based method of remote sensing to model the spatial distribution of organic carbon stocks in floodplain soils, using very high resolution remote sensing data and auxiliary data (e.g. Digital terrain model, historical maps). This method should support large-scale mapping of organic carbon stocks in floodplain soils. Knowledge from soil science can be regionalised and be used for the increased demands of landscape and environmental mapping, especially regarding the climate function of soils.
The methodology is based on spectral and knowledge-based classification. In order to determine vegetation parameters, such as vegetation type, age, density of trees per hectare, a combination of pixel and object-based classifications will be used. These results will be integrated into a rule-based classification process, using fuzzy logic to integrate additional data and expert knowledge into the spectral classification.


Quick Access

Schnellnavigation zur Seite über Nummerneingabe