Mapping and classification of Peatland on the Isle of Lewis using Landsat ETM

Britain contains some of the largest areas of blanket peatland in the world and the monitoring of this resource is vital. This study has investigated whether Landsat ETM+ can be used to identify types of blanket peatland on Lewis. This was done using Principal Component Analysis (PCA) on composites...

Full description

Saved in:
Bibliographic Details
Published in:Scottish geographical journal 2007-09, Vol.123 (3), p.173-192
Main Authors: Brown, E., Aitkenhead, M., Wright, R., Aalders, I. H.
Format: Article
Language:English
Subjects:
Classification
Environmental economics
Geographical information systems
GIS/cartography
land cover
Land economics
Landsat satellites
Mapping
Methodology
Peat
Peatlands
Principal components analysis
Remote sensing
Scotland
United Kingdom
Vegetation
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Britain contains some of the largest areas of blanket peatland in the world and the monitoring of this resource is vital. This study has investigated whether Landsat ETM+ can be used to identify types of blanket peatland on Lewis. This was done using Principal Component Analysis (PCA) on composites of band ratios and single band variables, and using neural network classification. The distinction between peatland and non-peatland was easily accomplished, but the identification of different peatland types was more difficult. PCA on a composite of spectral bands 1 to 9 was the most useful composite, but did not improve over the use of NDVI-related band ratios. An overlap was found between peatland classes caused by similar spectral signatures of peat banks and eroded peatland. This was confirmed by a separate study examining the variation between different blanket bog classes in the Land Cover of Scotland 1988 dataset. It is suspected that this problem will remain with Landsat ETM+ -based classification of peatland because the spatial resolution is insufficient to capture the heterogeneous nature of the terrain and vegetation types. The paper discusses further methodologies and information sources which, in combination with Landsat ETM+ data, could improve the ability to classify peatland.
ISSN:1470-2541
1751-665X
DOI:10.1080/14702540701786912