Study concludes forests are decreasing, but how good is the data?

Study concludes forests are decreasing, but how good is the data?

Using data from 541 randomly selected 18.5 by 18.5 km blocks, an analysis of Landsat-7 satellite data has concluded that 3% of the forests (1 million square kilometer) present in 2000 were gone by 2005.  See map at http://www.sciencenews.org/view/download/id/58673/name/WHEN_A_TREE_FALLS

The researchers conclude that the primary causes are wildfires and human-induced deforestation for agriculture and logging.  For purposes of their assessment, the researchers counted an area as "forest" if trees covered more than 25% of the landscape.  On a percentage basis, the largest loss (4%) was in the boreal region of the Arctic; fires caused 60% of the loss in this region, with logging accounting for much of the remainder.  Storms and beetle infestations (the latter being described in prior posts) also added to some of the remaining loss.

Unfortunately, the study did not measure the extent to which forests increased between 2000 and 2005, so recovery from previous natural disasters (e.g., wildfires, storms) was not included.  As such, it is difficult to assess the true net loss, if any.  Additionally, the researchers recognize that the Landsat data is not a perfect tool for this purpose.  The researchers have previously noted that sample design can have an important impact on the statistical validity of the conclusions to be derived.  See http://marte.dpi.inpe.br/col/dpi.inpe.br/sbsr@80/2008/11.18.00.55.25/doc/5671-5678.pdf.

The study can be found at http://www.pnas.org/content/107/20/9025.extract.