Scale in remote sensing and its impact on landscape ecology

Lechner, A 2010, Scale in remote sensing and its impact on landscape ecology, Doctor of Philosophy (PhD), Mathematical and Geospatial Sciences, RMIT University.


Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
lechner.pdf Thesis application/pdf 3.81MB
Title Scale in remote sensing and its impact on landscape ecology
Author(s) Lechner, A
Year 2010
Abstract The use of remotely sensed derived thematic data has become ubiquitous in landscape ecology. Remote sensing data has the potential to describe broad scale landscape patterns and relate them to ecological processes such as species persistence and distribution. However, these datasets are being used without considering the spatial uncertainty that is ever present in remote sensing data. Maps derived from remote sensing data will vary in the extent, patchiness and accuracy of their landcover classes predominantly due to the inter-relationships of a number of scale dependent factors such as pixel size, minimum mappable unit and categorical resolution. Furthermore, the effect of these factors on landcover classification is more pronounced in fragmented environments which are spatially complex, with habitat patches varying in size from roadside reserves (~10m2) to large vegetation remnants contained within national parks (100km2).

This thesis investigated the interaction and the relative importance of scale dependent factors on the characterisation of landscape pattern and ecological analysis using real and synthetic landcover data and simulated species-environment models. Several key findings emerged from this research. Firstly, it found that mapping error was highest when the scale of the feature and the raster grid coincided. Ecologically important landscape elements such as small and linear vegetation patches of similar scales to the raster grid had lower classification accuracies, and were less likely to be extracted than larger more compact features. Secondly, this thesis showed that at coarser scales, subtle levels of patchiness declined. Small patches either aggregated into larger patches or completely disappeared. Thirdly, it demonstrated that the ability to identify the scale at which a species interacts with the environment, using multi-scale species-environment models is affected by the scale of the remote sensing data. In conclusion, this thesis quantified the impact of scale on the classification of landcover maps and demonstrated how spatial uncertainty in the characterisation of landscape pattern can impact on ecological analysis. Without the incorporation of uncertainty arising from scale, ecological analyses using remote sensing data will continue to produce results with unquantified uncertainties, which may result in poor and/or ineffective management decisions.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Mathematical and Geospatial Sciences
Keyword(s) scale
spartial uncertainty
error
landscape pattern
remote sensing
land cover mapping
simulation modelling
Versions
Version Filter Type
Access Statistics: 517 Abstract Views, 1841 File Downloads  -  Detailed Statistics
Created: Mon, 06 Dec 2010, 11:10:27 EST
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us