ca-
ids
vas
the
> et
rig-
al-
)13C1
ed.
av-
ap.
fil-
this
vn-
wo
(2)
ing
1€
the
ing
ase
sti-
ach
|, a
sed
fol-
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B8, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
when K' reached first local maximum during the green-
up period, and End of Season (EOS) corresponded to
the time when K' reached last local minima during the
brown-down period. These moments were considered
as the transitions of vegetation growing from one lin-
ear stage to another (Zhang et al, 2003). The DOY
(Day of Year) of SOS and EOS started from July 1 of
each year. The length of season was calculated as the
difference between SOS and EOS. Other phenological
metrics were calculated using function fitted time se-
ries. Respectively, Maximum of Greenness (MaxG) and
Minimum of Greenness (MinG) were maximum and min-
imum EVI values during a phenological cycle; ampli-
tude (AMP) was the difference between MaxG and MinG;
Large Integral of Greenness (LIG) was the daily inte-
grated EVI value for a phenological cycle; Small Inte-
gral of Greenness (SIG) was the daily integrated EVI
during growing season (defined as a period between SOS
and EOS) subtracted daily integrated MinG during the
same period.
2.3.2 Segmented regression to identify breakpoints
in latitudinal gradients of phenological metrics In
this study, segmented linear regression was used to iden-
tify the breakpoint for phenological index along the lat-
itude. Segmented linear regression applies linear re-
gression to (xz, y) data that do not have a linear rela-
tion (Wayne Skaggs, 1996). On the relations between
phenological metrics and latitude, it is hypothesised that
there might be a significant breakpoint existed, however,
generally the relations do not necessarily have to be lin-
ear. So that with segmented linear regression, the break-
points can be introduced, among different segments, sep-
arate linear regression are applied and by this means
nonlinear relations between phenological metrics and
latitude might be approximated by a series linear seg-
ments (Wayne Skaggs, 1996). By calculation of the con-
fidence intervals of breakpoints, optimum breakpoint,
which means that the breakpoint with smallest interval
can be selected (Oosterbaan et al., 1990).
3 RESULTS AND DISCUSSION
3.1 Spatial patterns of vegetation phenology
In the NATT study area, all phenological metrics exhib-
ited significant spatial patterns in terms of 11 years of
average conditions (Fig. 2, 3,4, and 5). Over the NATT,
the dates of vegetation growing season onset (SOS) ranged
from approximately late August to late January, span-
ning a five months time period. While the dates of veg-
etation growing season dormancy (EOS) ranged from
late February to late October, spanning approximately
an eight months time period. The length of vegetation
growing season (LOS), which is the difference between
EOS and SOS, ranged from about 138 days to 354 days,
with differences as large as 216 days (7 months). Ta-
ble 1 and Table 2 provide detailed statistical summary
for all eight phenological metrics over the whole NATT
study area.
The latitudinal gradients of vegetation phenology were
also very significant in the NATT area. From north to
Table 1: Five-number summary 1 of phenological met-
rics in the NATT area. The phenological metrics were
the average of 11 years.
SOS EOS LOS MaxG
Minimum 50 243 138 0.1048
25% quantile 120 355 235 0.1794
Median 133 3384 298 02395
75% quantile 143 414 278 02921
Maximum 214 478 354 0.4902
south, the dates of vegetation growing season onset sig-
nificantly shifted to later by 3.916 days per latitude de-
gree (Fig. 6(a)).While the dates of vegetation growing
season dormancy also shifted south to later by approx-
imately 1.75 days per latitude degree (Fig. 6(b)), how-
ever, the latitudinal trend of EOS was not as strong as
the trend of SOS. The trends of SOS and EOS naturally
led to the latitudinal trend of LOS (length of season),
which showed a southward decreasing by approximately
1.6 days per latitude degree (Fig. 7(a)). The most sig-
nificant latitudinal gradient came from LIG (integral of
annual EVI), which can be considered as the vegetation
annual productivity, showed a almost straight line de-
cay, where the latitudinal averaged LIG in the south end
of NATT was only approximately 38% of the north end
(Fig. 7(b)).
In the temporal scale, the vegetation phenological met-
rics in the southern NATT generally had a relatively larger
interannual variabilities than the northern NATT (the ver-
tical lines in Fig. 6 and 7, which were the temporal stan-
dard deviations).
(a)
-12
Y
(Since July 1)
Latitude °s
1
=
128 130 Longitude vod 136 138
Figure 2: Spatial patterns of 11 years (2000-2011)
mean SOS (the date of growing season onset) in the
NATT.The numbers in the bracket indicate the calendar
years, which 1 means first half of phenological year, i.e.
from July 1 to December 31, 2 means second half of
phenological year, i.e. from next January 1. Original
0.05 degree resolution result had been aggregated to 0.2
degree resolution for plotting purpose.
3.2 Results of breakpoint analysis
The breakpoint analysis showed that at least in terms
of annual minimum EVI (MinG), which was considered
as a good indicator of tree cover ratio, there was a de-
tectable change around 18.84 °S and 20.02 °S (Fig. 8),
