Bock, M. and Krause, G. (2006) Remote Sensing of Mangrove Forest Structure and Dynamics. MADAM . p. 7.

[img] Text
Bock 2006.pdf - Published Version
Restricted to Registered users only

Download (990kB)


On a mangrove peninsula in North Brazil suitable and transferable techniques for the remote sensing and airborne-based classification for the assessment of mangrove forest structures were developed within the framework of the MADAM (Mangrove Dynamics and Management) Project. Examples of this research are presented in this article. On local scale an object based classification rule network, which was developed and applied for airborne photo-transects, was adapted to Ikonos sub-scene. The transfer of the object based rule network to the whole peninsula however, exposed crucial deficiencies in the overall technical performance. On this level, the rule network was redesign to fuzzy sets instead of using training sets and nearest neighbour algorithms. The results provide an impressive insight into the structure of mangrove ecosystems and their hinterlands. Change Detection On the landscape level about 7 Landsat TM and ETM scenes ranging from 1986 to 2003 served as a remote sensing database to analyse the performance of several change detection techniques. These proved to be unstable and lead to the development of a new methodological approach based on land cover signature specific conditional queries. It was successfully applied to monitor the dynamics of costal mangroves covered by four selected images from 1986 to 2001. These are currently compared with field data of the long-term beach profile monitoring programme. Moreover three Corona spy-satellite black & white photos with a resolution of about 4m from 1967-1969 were employed. As these scenes display the mangrove peninsula as an undisturbed ecosystem prior to the road construction in the 70ies, the detection of major changes compared to the Ikonos scene of 2003 were possible.

Document Type: Article
Programme Area: UNSPECIFIED
Research affiliation:
Refereed: Yes
Open Access Journal?: No
Date Deposited: 18 May 2021 20:04
Last Modified: 18 May 2021 20:04

Actions (login required)

View Item View Item