Geoscience Reference
In-Depth Information
8.3.2 Rates of Change and Sources of Imperviousness (1992-2005)
A portion of the USGS's National Land Cover Dataset (NLCD) 1992/2001 Retrofit
Land Cover Change Product (LCC9201) (MRLC
2009
; Fry et al.
2009
)was
extracted and geoprocessed to the extent and boundary of Kentucky. In particular,
three “from-to” - or “source-target” classes that are inherent in the LCC9201 data
were extracted which included those labeled
“62”
(from agriculture to developed),
“42”
(from forest land to developed), and
“12”
(from water to developed). These
classes represented 74.3, 23.4, and 1.9% of the added development, respectively. A
“boundary clean” algorithm was utilized to eliminate single pixel change locations.
A negative consequence of this operation is the loss of some linear change features,
and some areas of less dense development (Fig.
8.8
).
Fig. 8.8
Retention and loss of regions of change; (
a
) compact, (
b
) linear and (
c
) less dense and,
respectively due to a generalization or “boundary clean” operation. Retrospective change (1992-
