Environmental Engineering Reference
In-Depth Information
IUTI spatial layout of sampling points follows a tessellation stratified sampling
design (also known as unaligned systematic sampling); this sampling scheme
is preferable to simple random or systematic grid sampling as sample plots are
distributed evenly to all parts of the target area according to a random scheme
(Barabesi and Franceschi
2011
). The set of sample points is extracted using a
0.5 km square grid, for a total of about 1,206,000 geo-referenced points randomly
located in each square cell and fully covering the Italian territory (Fig.
4.1
). A sub-
set of the IUTI sample is represented by the 301,300 first phase sample points of
the INFC.
IUTI sampling interval and the large sample size are designed to capture and
estimate land use changes from 1990 to the end commitment period with a suit-
able statistical accuracy. Each sample point is photo-interpreted on the basis of
digital orthophotos and classified into GPG-LULUCF derived land use classes at
given points in time (1990, 2008, 2012). For sample points where land use change
in the forest category is detected between 1990 and 2008, as result of AF/R and D
activities, the land use classification is performed also at an intermediate point in
time (2000), in order to estimate, by interpolation, the annual gain/loss of forest
area in different time periods (1990-2000, 2000-2008, 2008-2012). The classifi-
cation of the sample set is currently completed for the years 1990, 2000 and 2008.
The estimation of the area of each land use class at national level, and its break-
down by administrative districts, is performed via area proportions, based on the
methodology reported by Corona et al. (
2012
).
Fig. 4.1
Exemplification of
IUTI tessellation stratified
sampling scheme; in the
background, boundaries of
Italy administrative units
(district level)
