The World Resources Institute blog has a fascinating article about the greening of Ethiopia in recent years. This is largely based on a paper in Science of the Total Environment by Nyssen et al.
Jan Nyssen, a geographer at the University of Ghent, together with a large team of colleagues reviewed historic photographs of the Ethiopian highlands comparing them to more recent pictures of the same locations. They then assessed the change in greenery in the landscape, using a benchmarking methodology to assign a value, before performing a multiple regression analysis against possible causes.
The upshot of their paper is that these welcome changes seem to be largely the result of human activity - the landscape has greened the most in areas of higher population density. This seems to be related to factors such as the planting of eucalyptus trees in response to market demand for poles as well as more obvious conservation measures such as the use of terracing to prevent water run-off and soil erosion.
I'm slightly bemused by the authors' conclusion though (my emphasis).
We conclude that except for an apparent upward movement of the upper tree limit, the direct human impacts on the environment are overriding the effects of climate change in the north Ethiopian highlands and that the northern Ethiopian highlands are currently greener than at any other time in the last 145 years.
The problem is apparent when you refer to the graph of their correlation analysis:
The red line is the rainfall graph, precipitation presumably being the key climatic variable in such an arid area. How can human activity be "overriding" the effects of climate change when any such changes have, if anything, been entirely beneficial?