Figure 1. Gain, loss and balance of changes for each class. The bars represent the percent change relative to the total change area (259,285 km
2
) with the
equivalent area in square kilometers annotated at the end of each bar. The colors indicate the proportional contribution from which class area was gained
and to which class it was lost. Bars with red outline correspond to the scale in red.
and Mexico were omitted. For instance, in comparison to
10m Système Pour l’Observation de la Terre (SPOT) 5 High
Resolution Geometrical (HRG) imagery Colditz et al. (2014)
calculated an omission error of 51% for potential change areas
in Mexico; and Fry et al. (2011) demonstrated with Landsat
5 TM and 7 ETM
+
data that the highest U.S. change rates
spatially occur in the Southeast U.S. – a pattern not captured
at the 250m scale.
Class-specific Change Between 2005 and 2010
The following section represents for each class the total area
gained and lost between the maps of 2005 and 2010 and the
balance as the difference between gain and loss. Numbers in
percent for gain, loss and balance in Table 2 are relative to
the total change area (259,285 km
2
); and Figure 1 visualizes
the changes including the contribution of each class.
Class Temperate or sub-polar needleleaf forest (15.06% of
the total area), which dominates in boreal Canada, the western
Cordilleras from Alaska to mountainous areas in Mexico, and
the southeastern United States, showed the highest negative
balance (-29.47%) mostly because of forest fires and timber har-
vest. Substantial class-specific net losses were to Mixed forest,
Temperate or sub-polar shrubland, and Temperate or sub-polar
grassland. Located in a similar environment in boreal Canada,
Forest harvesting is another agent of
change as well as insect-induced mortality.
The mountain pine beetle infestation is
responsible for much of the change observed
in southern British Columbia.
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
October 2014
921
