Quantifying the spatial and temporal changes in forested landcover using landscape metrics derived from remotely sensed data in rural parts of Zimbabwe

Abstract

This study examines the application of remotely sensed derived landscape metrics to understand the changes and dynamics in the form and morphology of forested areas in rural parts of Zimbabwe from 2002 to 2011. Specifically, the study determines the spatial and temporal changes in forest areas due to human activities, such as crop production, using landscape metrics derived from classified Landsat remote sensing images. The results from this study have shown that landscape metrics derived from the 30-m Landsat dataset have a great potential for understanding the patterns of change in forested areas in the developing world. For instance, in the year 2002, the majority of the land was occupied by forests, whereas in 2011, it was shown that non-forested areas became more dominant and scattered, and forested areas showed a decrease in spatial extent. Moreover, statistical results have shown that in 2002, the number of forest patches was higher and significantly different when compared to non-forested patches. However, in 2011, both land cover types showed an increase in the number of patches, although the number of non-forested patches was higher and indicated significant differences (p < 0.05) in terms of areal extent. Overall, the findings of this study provide a comprehensive understanding of the impacts of human activities on the natural ecosystem which leads to better and improved future land use planning and management strategies.

Keywords:

• croplands
• fragmentation
• remote sensing
• landscape metrics
• spatial variations

ACKNOWLEDGEMENTS

The authors would like to express their gratitude to the United States Geological Survey (USGS) for providing Landsat images and anonymous reviewers for their contribution.

ORCID

Timothy Dube http://orcid.org/0000-0003-3456-8991