rubber plantations were already present, exhibiting a circle-
layer pattern around Jinghong City. The other, but more
important trend is a continual expansion toward the border
regions with Myanmar and Laos. Spatially, the expansions
varied from concentrated to dispersed, gradually expanded
into border regions to the south (i.e., northern Shan in
Myanmar and Luang Namtha Province in Laos) and east (i.e.,
Phongsaly Province in Laos). As geo-economic cooperation
has gradually increased since the 1990s, we noticed during
our field surveys that deciduous rubber plantations have ex-
panded out of Xishuangbanna and the border regions between
China, Laos, and Myanmar. Taking a 30-km buffer along the
border as an example, the area of deciduous rubber planta-
tions within the China-Laos border region on the Chinese
side were 171, 177, 525, 965, and 1334 km
2
, respectively, in
the early 1990s, late 1990s, early 2000s, late 2000s, and early
2010s—an overall 7.8-fold increase of 1163 km
2
. Meanwhile,
the area of deciduous rubber plantations within the China-
Myanmar border region on the Chin
about 7.5 times (1366 km
2
) from 21
306 km
2
by the late 1990s, 640 km
2
km
2
in the late 2000s, and 1577 km
2
Vertical Expansion of Deciduous Rubber Plantations During 1991–2016
In addition to the marked horizontal expansion of decidu-
ous rubber plantations revealed, the data also indicate a clear
vertical expansion over time with increments of altitude.
Utilizing 30-m
ASTER GDEM
data, we analyzed the relationship
between the distribution of deciduous rubber plantations in
recent decades and elevations (Figure 10).
In Figure 10, deciduous rubber plantations were relatively
concentrated at elevations ranging between 600 and 1000
m across five past epochs: 88.2% in the early 1990s, 83.4%
in the late 1990s, 88.4% in the early 2000s, 87.7% in the
late 2000s, and 94.1%, in the early 2010s. Slight differences
are nevertheless seen in the elevations at which the largest
proportions of plantations were present over the last 26 years.
During the early 1990s, the deciduous rubber plantations
at elevations between 600 and 700 m comprised the largest
proportion (53.3%, 243 km
2
), while less than 1.2% of the total
area was at elevations less than 500 m or more than 1200 m.
During the late 1990s, the peak at elevations between 600
and 700 m was 30.2% (159 km
2
) of the total deciduous rubber
plantations. This changed by the early 2000s, as the elevation
band containing the largest proportion of deciduous rubber
plantations rose to between 700 and 800 m, encompassing an
area of 372 km
2
(32.1%) in the early 2000s, 635 km
2
(30.4%)
in the late 2000s, and 825 km
2
(28.8%) in the early 2010s,
while just 1.3%, 0.7%, and 0.3%, respectively, occurred at
elevations below 500 m or above 1200 m in the corresponding
epochs.
Deciduous rubber plantations within Xishuangbanna have
expanded toward both higher (over 1000 m) and lower (below
600 m) elevations since 1991. It is also noteworthy that these
plantations broke through their previous upper elevation lim-
its of about 1000 m in the early 1990s, about 1100 m in the late
1990s, about 1200 m in the late 2000s, and about 1300 m in the
early 2010s, and also expanded to their lower limit of around
500 m in the early 2010s. These changes indicate that the
expansion of deciduous rubber plantations was most marked
in both higher- and lower-elevation areas because of higher
profitability. The maps also indicate that during this period,
blished deciduous rubber planta-
tions between 800 and 1100 m; the
35-fold from 40.03 km
2
to 1350.56
hese data indicate that plantation
sed by new establishment at higher
and lower elevations. Although there was a clear expansion
trend towards more elevated areas due to higher economic
cost and unfavorable biophysical conditions (i.e., water and
temperature), the increments of deciduous rubber plantations
have been primarily limited below 1200 m, with only 0.2%
(6.15 km
2
) above this level. Historical horizontal and vertical
expansion data clearly suggest that future suitable regions for
rubber plantation are limited in Xishuangbanna.
Discussion
Apart from reconstructing the change trajectories of 30-m-
resolution deciduous rubber plantations, this study might
have two critical implications. First, the reliability of
NBR
in
detecting rubber plantations was reverified in this study. In
our previous study, Landsat-derived
NBR
was proven more
sensitive to detecting deciduous rubber plantations (Li
et al.
2015). In this updated study, the tri-window
CRNBR
method
held greater potential in tracing back the past expansion in-
formation of deciduous rubber plantations. An improvement
of our approach was the introduction of dense phenological
Figure 10. The area and proportion of deciduous rubber plantations at different elevation gradients within Xishuangbanna.
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
September 2019
695
