Conference Contribution Details
Mandatory Fields
Ali, A., de Bie, C.A.J.M., Skidmore, A.K., Scarrott, R.G.
2nd TERRABITES Symposium
A new landscape heterogeneity method for accurately mapping natural and semi natural landscape
ESRIN, Frascati, Italy
Poster Presentation
Optional Fields
Natural and semi-natural landscapes are often heterogeneous in terms of landcover. Mapping landcover requires an approach that considers both gradient representation and landcover long term spatio-temporal variability. These aspects can be visualized and depicted by applying a new spatio-temporal analysis based Landscape Heterogeneity Mapping (LaHMa) method to natural and semi natural landscape. Using MODIS NDVI 16-day imagery (Feb2000-Jul2009) for Crete, the image having sixty five (65) clusters were selected from ISODATA classification results using high average and minimum separability values of divergence statistics. The 65 clusters image represents the optimal number of clusters to appropriately generalize the spatial and temporal variability in landcover of that area. This indicates the number of cluster images to be used for generating landscape heterogeneity map. Using classified outputs from 10 to 65 clusters, the frequency of pixels declared as boundary of homogeneous landcover classes were translated in the form of landscape heterogeneity map. The map was validated using field data. The results showed that heterogeneity map expressed as boundary strength showed moderate correlation (R2 = 0.60) with sum of differences between neighbor transect pixels in all landcover components (trees, shrubs, grass, bare soil, stone and litter cover). Generally the study found this new approach (LaHMa) suitable for mapping landscape heterogeneity in natural and semi-natural landscape of Crete, Greece. The method can be particularly useful for gradient analysts and landscape ecologist in term of heterogeneity in ecology.