To assess forest loss, we followed Hansen et al. (2013) methodology, and defined forest cover as vegetation taller than 5 m and forest loss as the complete removal of tree canopy at a 30 m resolution. Hansen forest-cover change data was extracted and processed in the Google Earth Engine. We only looked at forest cover loss not gain, which was not included in this analysis for two reasons: young forests are unlikely to support forest-dependant species, and much of the gain can be attributed to monoculture plantations of oil palm or rubber which are major threats to tropical forests. There are limitations of satellite-derived estimates of global forest change, such as an inability to differentiate between ecologically valuable forest and agro-forests, such as oil palm, and lower accuracy in more arid environments. Likewise, ground truthing is required to infer the causes of forest loss since the dataset does not differentiate between ecologically harmful clearing, and purposeful clearing for example of invasive species, which has a conservation benefit. But even with these limitations, the Hansen et al. (2013) forest data product is considered the most accurate global representation of temporal loss of forest available.
Details about the work are provided in the following paper:
What does it mean for natural World Heritage sites?
We quantified forest loss by year in each natural World Heritage site and plotted against forest cover in 2000 (y axis) - this shows the amount of forest loss that has happened and highlights the year where major forest loss events occurred. The interactive map can be used to display cumulatively forest loss from 2000 onwards.
While possible to estimate the amount of forest loss, it is difficult without ground truthing data to ascertain the reason behind the loss. Nevertheless, it could be employed to help identify sites that may merit further investigation.