Classification and use of landform information to increase the accuracy of land condition monitoring in Western Australian pastoral rangelands
Historically landforms have been mapped from field-based surveys or using stereo aerial photographs. Information about landforms plays an integral role for landscape evaluations, suitability studies, erosion studies, hazard predictions and various fields of landscape and regional planning. First, te...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Conference Paper |
| Published: |
WALI
2011
|
| Online Access: | http://www.walis.wa.gov.au/isde7/peer-reviewed-papers/classi-1.pdf/download http://hdl.handle.net/20.500.11937/25799 |
| Summary: | Historically landforms have been mapped from field-based surveys or using stereo aerial photographs. Information about landforms plays an integral role for landscape evaluations, suitability studies, erosion studies, hazard predictions and various fields of landscape and regional planning. First, techniques and software were explored to extract primary topographic attributes from a digital elevation model (DEM): elevation, slope, profile curvature and plan curvature. Second, LandSerf software was employed to classify the DEM into landform types through semi-automated feature extraction. Six landform classes were produced: pits, channels, passes, ridges, peaks and planar. Experimental methods established the most suitable sampling window scale and slope tolerance. Curvature was not used in the classification process therefore no differentiation was found between slope types: upper, mid, and lower slope. Object based image analysis was tested using Landsat TM imagery. The imagery was segmented into areas with similar shapes. The boundaries can be forced to replicate those of landscape variables including the landform data produced in LandSerf, land system boundaries and available land unit boundaries. This research provides evidence that available software can be used to map landform elements at the land subsystem level where currently data only exists at a land system level. This research aims to increase the quality and quantity of available land condition data that can be used in monitoring conditions of pastoral leases. Greater ability to provide accurate results on pastoral conditions will enable the lessee to better manage their land and increase productivity. |
|---|