Understanding patterns of vegetation change at the Angkor World Heritage site by combining remote sensing results with local knowledge

ABSTRACT

This research uses two disparate data types, quantitative bi-temporal Multivariate Alteration Detection (MAD) remote sensing with qualitative interview data, to examine the spatial and temporal relationships between patterns of vegetation change at the Angkor World Heritage site, Cambodia. The time period of the research data corresponds with periods where there were significant shifts in the regulations affecting management of the site. The findings suggest that components of change identified from MAD across different spatial and temporal scales can be translated to that observed in the field or ‘real world’. Although MAD results provide an indication of change at specific locations, the results from field investigations and interviews suggest, however, that the different forms of change identified are not always clearly understood. For example, in the case where MAD outputs and associated spectral plots suggest evidence of fire in the landscape, it is not clear whether this is the main contributor to vegetation loss at given sites. The research demonstrates that identifying patterns of vegetation change from satellite imagery is achievable using novel and largely under-utilized methods. However, understanding the forms of change and contributing factors are considerably more complex.

Acknowledgements

I thank the communities of Angkor who participated in this research. Appropriate ethical standards approved for this research were followed at all times.

All figures by the author, except where otherwise indicated.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. A concept adopted by UNESCO to describe the interaction between culture and nature in the context of World Heritage.

2. Landsat images were obtained from the United States Geological Survey (USGS) Earth Resources and Science Centre (EROS) (http://glovis.usgs.gov). ASTER images were obtained from the Earth Remote Sensing Data Analysis Centre (ERSDAC) Japan (http://www.ersdac.or.jp/eng/index.E.html).

3. At the time of image acquisition only bands 1, 2 and 3 were freely available, hence no additional ASTER bands were used.

4. A 2% linear histogram colour stretch was applied to all change images for display purposes.

5. Identified as regions of interest (ROIs) in ENVI™.

6. Location of villages, and their proximity to change categories can be seen in Figure 8.

7. Forest that has experienced successional change a long time in the past.

8. A complete list of trees identified during the course of this research are available directly from the author. Text in brackets are plant names in the Khmer language.

9. Swidden is a form of chamkar agriculture which is typically not permanent.

10. Location of villages, and their proximity to change categories can be seen in Figure 8.

Additional information

Funding

This work was supported by the ARC-Linkage Project: ‘Living with Heritage: Integrating Time, Place and Culture for World Heritage Conservation’ [LP0454989].