Per-pixel, post-classification comparison results indicate that approximately
26% of the study area exhibits LCLU change for the study
period. A summary of LCLUC transitions in terms of the percent of
total land cover change between 2000 and 2010 is presented in
Table 7. As with areal estimates provided in Table 6, change transition
percentages were adjusted to correct omission/commission errors identified through the accuracy assessment of 2000–2010 LCLU changes
(Table 4). Land cover transitioning to Built included Agriculture (6%, primarily
around the periphery of the Accra and Kumasi metropolitan
areas) and Natural Vegetation (1%). The highest percentage (69%) of
LCLUC is associated with conversion of Natural Vegetation (including
Forest, Secondary Forest, and Savanna) to Agriculture. The next highest
percentage (25%) of LCLUC is associatedwith transition of Agriculture to
Natural Vegetation between 2000 and 2010.Most of this change is associated
with transition from Agriculture to Savanna/Low Stature Natural
Vegetation.We explored potential data issues that might negatively impact
classification results, but found none. This included analyzing patterns
of no-data gaps to see if there was any correspondence,
examining and ensuring correct cloud screening, and examining the
dates of imagery (relative to the growing season) to see if any differences
in timing between 2000 and 2010 could explain this.We also examined
2010 high spatial resolution imagery for limited areas where
this transition occurred. Review of the image data utilized for classification
and the 2010 high spatial resolution imagery indicate that
large portions of land in the Afram Plain were allowed to go fallow
during the study period. Further, field work performed in Ghana
in September 2015 supports this conclusion, as 40 out of 48 sample
sites (83%) that exhibited this transition show evidence of former
agriculture lands (including tree plantations) in a fallow or abandoned
state and mostly replaced by savanna grasses. However, while the assumption
is that the fallow land will return to agriculture, it is possible
that these lands have been abandoned. Continued investigation is
required.
5. Discussion