Download The Three-North Shelterbelt Program and Dynamic Changes in

March, 2014
Journal of Resources and Ecology
J. Resour. Ecol. 2014 5 (1) 053-059
DOI:10.5814/j.issn.1674-764x.2014.01.006
www.jorae.cn
Vol.5 No.1
Article
The Three-North Shelterbelt Program and Dynamic Changes
in Vegetation Cover
WANG Qiang1, ZHANG Bo2*, ZHANG Zhiqiang1, ZHANG Xifeng3 and DAI Shengpei4
1 Scientific Information Center for Resources and Environment, Lanzhou Branch of the National Science Library, CAS, Lanzhou 730000, China;
2 College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, China;
3 Center for Dryland Water Resources Research and Watershed Science, Key Laboratory of West China’s Environmental System (Ministry of Education),
Lanzhou University, Lanzhou 730000, China;
4 Institute of Scientific and Technical Information, Chinese Academy of Tropical Agricultural Sciences, Danzhou 571737, China
Abstract: The Shelterbelt Forest System Program in northeast, north and northwest China (the Three-North
Shelterbelt Program, TNSP) is the largest ecological reforestation program in the world. TNSP vegetation research
has important ecological meaning and profound social and economic significance. Here, spatio-temporal variation
in vegetation cover under the TNSP was examined using the NDVI average method, major climatic factors such
as temperature and precipitation, and linear regression trend analysis from 1982 to 2006. We found that in the
past 25 years, NDVI vegetation in the study area has consistently risen at a rate of 0.007 per decade. Vegetation
cover, temperature and precipitation are positively correlated. The area of vegetation associated with precipitation
is larger than the area related to temperature; precipitation is the key factor affecting vegetation growth across the
TNSP. From 1982 to 2006, regions with improved vegetation cover were found in the central and southern part of
the Greater Khingan Mountains, central part of the Lesser Khingan Mountains, northeastern part of the Changbai
Mountains, Yanshan Mountians, Western Liaoning Hilly Region, Altai Mountains, Tien Shan Mountains, eastern
part of the Qilian Mountains, eastern part of the northwest desert as well as southern part of the Gully Region of the
Loess Plateau. However, vegetation cover declined on both sides of the Greater Khingan Mountains, western part
of the Hulun Buir Plateau, northern part of the Sanjiang Plain, southern part of Horqin Sandy Land, southern part of
the northwest desert and northern part of the Gully Region of the Loess Plateau.
Key words: NDVI; Three-North Shelterbelt Program (TNSP); spatio-temporal changes; correlation analysis;
dynamic study; northern China
1 Introduction
As the main component of the terrestrial ecosystem, plants
are sensitive indicators of climate and human factors, and
tie together the soil, air and water. Plants play a regulation
and indicator role in energy balance, biological chemical
cycles and water circulation. Changes in plants are the
product of the earth’s interior (e.g. soil parent material and
soil types) and external (e.g. temperature, precipitation and
human activity) environments (Kaufman and Tanre 1992;
Carlson and Ripley 1997; Ichii et al. 2002; Jason and Roland
2004; Pettorell et al. 2005; Song and Ma 2008). At a large
scale and mesoscale, remote sensing is an effective means
for monitoring global and regional vegetation (Adams and
Arkin 1997; Shi et al. 2006). The normalized difference
vegetation index (NDVI) is the most widely used vegetation
monitoring data via remote sensing, and indicates changes
in vegetation cover, biomass and ecological systems (Tucker
and Townshend 1985; Defries and Townshond 1994; Tucker
et al. 2005; Lasaponara 2006; Brown et al. 2008; Wang et al.
2011). NDVI is a ratio parameter of near infrared band and
infrared band using the formula:
NDVI = (NIR-VIS) / (NIR+VIS)
Received: 2013-06-09 Accepted: 2014-03-03
Foundation: National Natural Science Foundation of China (41171116; 40961038).
* Corresponding author: ZHANG Bo. Email: [email protected].
54
Journal of Resources and Ecology Vol.5 No.1, 2014
where, NIR indicates near infrared band, and VIS indicates
infrared band. NDVI can accurately reflect vegetation
greenness, photosynthesis intensity, vegetation metabolic
intensity and its seasonal and interannual variability.
The Three-North Shelterbelt Program (TNSP) is a
major ecological construction project underway in China.
This project will transform nature at the largest scale and
over the longest time span in human history. Vegetation
dynamics resulting from the TNSP will have important
ecological environmental significance and profound social
and ecological significance. Due to the dramatic effect of
long-term human activities and global climate warming,
problems such as flood disasters, desertification and soil
erosion are becoming more serious in northern China. For
example, the area affected by desertification has reached
1.49 million km2 (Zhu et al. 2004). Based on a 1982-2006
GIMMS NDVI data set, statistical methods and vegetation
cover mapping, we analyzed temporal and spatial variation
and characteristics of vegetation under the TNSP.
3 Data acquisition and methodology
2 Study area
Monthly NDVI was obtained by 15-day composed NDVI
data with maximum value composed (MVC), i.e. the
monthly NDVI image pixel was replaced by the maximum
NDVI values of two 15-day composed NDVI values, which
can reduce the negative influence of atmospheric clouds,
particles, shadows, perspective and solar altitude (Stow et
al. 2007). Then, the average method was used to composite
the annual average NDVI, which was representative of the
most abundant vegetation period within a year.
75˚0΄0˝E
90˚0΄0˝E
105˚0΄0˝E
120˚0΄0˝E
135˚0΄0˝E
N
GIMMS NDVI data from 1 January 1982 to 12 December
2006 was based on NOAA meteorological satellite global
8 km × 8 km data issued formally by NASA and widely
used in global vegetation change research. The 15-day
NDVI maximum synthetic data are generated by radiation
correction, geometric correction, atmospheric correction
and reflectivity calculation; spatial resolution is 8 km
with Albert equal-area conic projection. The DN value is
converted to NDVI using the formula NDVI = 0.008×(DN–
128) by ArcGIS 9.2 and the NDVI time series for 1982–
2006 was clipped by boundary vector data. Land cover
data from the global land cover data subset of China was
developed by CLC2000. Precipitation data was provided by
the National Meteorological Library.
3.2 Methodology
3.2.1 Maximum value composited of yearly NDVI
Y =
1
25
25
∑ NDVI , i=1, 2, …, 25
where, Y is composited the total annual average NDVI;
NDVI is the annual average NDVI from 1982 to 2006; i is
1 for 1982, 2 for 1983 and so on.
3.2.2 Trend line analysis
Simulating the trend of each raster grid reflects the spatial
40˚0΄0˝N
Provincial region
I Sandy area
II Northwest desert area
III Gully region Loss Plateau
IV Northeast North China plain
30˚0΄0˝N
30˚0΄0˝N
40˚0΄0˝N
Legend
90˚0΄0˝E
105˚0΄0˝E
120˚0΄0˝E
(1)
i =1
50˚0΄0˝N
50˚0΄0˝N
Construction of the Three-North Shelterbelt started in 1978
and is predicted to take 73 years to complete. Altogether,
eight stages will be needed to build a forest shelterbelt
along the Great Wall of China. The objective of the TNSP
is to control sand and wind erosion, harness soil and water
losses, improve ecological environments and produce
multiple forest products. The project is acclaimed as China’s
Green Great Wall and World’s Best Ecological Project both
nationally and internationally (Li 1994). The area covers
551 counties and 13 provinces (73°26′E to 137°50′E;
33°30′N to 50°12′N), covering an area of 4.069×106 km2 or
42.4% of China (Fig. 1). The west is higher than the east,
and the altitude ranges from 100–5000 m a.s.l. According to
the natural climate, geomorphic features, fragile ecological
environment and regional otherness of economic levels, the
area can be divided into four areas: the Sand, Northeastern,
North China Plains and Northwestern Desert Areas.
3.1 Data acquisition
Fig. 1 The Three-North
Shelterbelt Program area.
55
WANG Qiang, et al.: The Three-North Shelterbelt Program and Dynamic Changes in Vegetation Cover
characteristics of changes in vegetation cover (Stow et
al. 2003; Ma et al. 2006). The change in trend can be
calculated for a certain period and respective pixel by the
linear regression of one variable, y=ax+b. Here, x is the year
number from 1 to 10, and y is the MNDVIi value for each year.
We can use (2) to calculate the slope of the linear time trend
by OLS (ordinary least squares) estimation:
n×
θslop=
n
∑i× M
i =1
NDVI i
−
n
n
∑ i∑ M
i =1
i =1
 n 
n∑ i −  ∑ i 
 i =1 
i =1
n
NDVI i
(2)
2
2
where, n is cumulative number of monitoring years; and θ
slope is the slope of the linear regression of one variable
equation. It is the average annual increase (or decrease) of
NDVI from 1982 to 2006.
3.2.3 Correlation analysis of vegetation and climate factors
Correlation analysis is commonly applied to analyze the
relationship between inter-annual vegetation changes and
climatic factors. Here, study area NDVI, annual mean
temperature and annual precipitation for each pixel was
N
Legend
0
500 1000
km
High: 1
Low: -1
Fig. 2 Spatial distribution of NDVI in the Three-North Region
of China from 1982 to 2006.
N
used in spatial correlation analyses to reflect the correlation
between climatic factors and NDVI sequences using
ArcGIS. The correlation coefficient between –1 and 1, was
calculated as follows:
∑ (x
n
rxy =
i
i =1
∑ (x
n
i =1
i
−x
−x
)(y − y )
i
) ∑ (y − y )
2
n
i =1
(3)
2
i
where, n is the cumulative number of monitoring years; x
and y are two variables related to the analysis; and x and y
are the variable sample mean value.
4 Results and discussion
4.1 Spatial distribution characteristics of vegetation
cover
Distribution characteristics of the NDVI mean value
show spatial variation. According to Fig. 2, the general
characteristics of vegetation cover in the three-north
region show the order IV area>III area>II area>I area; the
vegetation cover status of the northeast and north plain
regions is better than that over the northwest and sandy
regions.
The maximum NDVI value is in the IV area and the
value in most areas is more than 0.75. The forest types
are evergreen, deciduous coniferous forest broad-leaved
deciduous forest and shrub located in northern Greater
Khingan Range, southern-central Lesser Khingan Range,
Yanshan Mountains, Changbai Mountains, Sungari River
Basin and Haihe River Basin (Fig. 3). NDVI values of the
III and I areas are second place, and the value in most of
these areas is in the range of 0.5–0.75. The forest types are
evergreen, deciduous coniferous forest and broad-leaved
deciduous forest mostly distribute in south-central Greater
Khingan Range, Yinshan Mountains, Qinling Mountains
and the Yellow River Basin. The minimum NDVI value is
in the II area and the value of most areas is less than 0.5.
In a few regions the value is in the range of 0.6–0.8 and
located in the Tien Shan Mountains, west-central Kunlun
Mountains, natural oasis in Xinjiang, east of the Qilian
Mountains, Qaidam Basin and Tarim Basin. The forest types
are evergreen and deciduous coniferous forest. Forestland
had the highest NDVI value.
4.2 Temporal characteristics of vegetation cover
0
500 1000
km
Legend
Deciduous needle-leaf forest
Evergreen coniferous forest
Deciduous broad leaved forest
Shrub
Fig. 3 Spatial distribution of forest in the Three- North
Region of China.
Analyzing the 25-year mean NDVI value curve we found
that the minimum NDVI value was 0.584 appearing in 1982
and the maximum value was 0.634 occurring in 2003 (Fig.4).
The annual average linear change trend was 0.007 per 10
years which suggests that the NDVI value has followed a
general slow upward trend; the correlation coefficient of
annual NDVI value and year is 0.40 (p < 0.05). Specifically,
the minimum NDVI value in the sandy, northwest desert,
Loess Plateau and northeast-north China plain areas was
56
Journal of Resources and Ecology Vol.5 No.1, 2014
I
0.95
II
III
y=0.0016x–2.2263
R2=0.8216
0.86
0.77
IV
The study area
y=0.0008x–0.9327
R2=0.1475
0.68
0.59
y=0.0025x–4.2955
R2=0.3859
y=–0.0002x–0.8489
R2=0.0103
y=0.0007x–0.8422
R2=0.177
0.50
0.41
1982
1985
1988
1991
1994
1997
2000
2003 2006
Fig. 4 Inter-annual variation of the NDVI in the TNSP.
0.636 in 1989, 0.428 in 1995, 0.928 in 1982 and 0.935
in 1982, respectively. The maximum NDVI value in the
sandy, Loess Plateau, and northeast-north China plain areas
was 0.748, 0.768 and 0.976, respectively, all in 2003. In
the northwest desert area the maximum value was 0.479
in 1988. The NDVI value increased most in the sandy and
northwest desert areas (p < 0.01) with a linear change trend
of 0.025 and 0.016 per 10 years, respectively. The Loess
Plateau (p < 0.05) had a linear change trend of 0.008 per 10
years. The northwest desert area showed a low decreasing
trend (p < 0.05) with a linear change of –0.002 per 10 years.
4.3 Spatial change characteristics in vegetation cover
Table 1 and Fig. 5 show vegetation cover patterns for 1982
to 2006, 1982 to 1991, 1991 to 2001, and 2002 to 2006.
Vegetation cover change across the TNSP area is increasing.
Table 1 NDVI changes in the Three-North Region of China.
Rank
1982–1991
1992–2001
2002–2006
1982–2006
θ slope
Change level
Area (km2) Ratio (%) Area (km2) Ratio (%) Area (km2) Ratio (%) Area (km2) Ratio (%)
< –0.005
Significant reduction
23 407
0.58
27 453
0.67
17 627
0.43
15 462
0.38
–0.005– –0.002 Slightly reduced
174 310
4.28
242 739
5.96
165 063
4.06
166 015
4.08
–0.002–0.002 Basically unchanged 3 398 289
83.52
3 347 776
82.28
3 451 114
84.81
3 338 207
82.04
0.002–0.005
Slightly increase
338 100
8.31
355 092
8.73
372 546
9.16
441 893
10.86
> 0.005
Significant increase
134 894
3.31
95 940
2.36
62 650
1.54
107 423
2.64
N
0
N
a
Legend
Significant reduction
Slightly reduced
Basically unchanged
Slightly increase
Significant increase
500 1000
km
N
0
b
500 1000
km
N
c
d
Legend
0
500 1000
km
Significant reduction
Slightly reduced
Basically unchanged
Slightly increase
Significant increase
Legend
Significant reduction
Slightly reduced
Basically unchanged
Slightly increase
Significant increase
0
500 1000
km
(a: 1982–2006; b: 1982–1991; c: 1992–2001; d: 2002–2006)
Fig. 5 Spatial distributions of NDVI in the Three-North Region of China.
Legend
Significant reduction
Slightly reduced
Basically unchanged
Slightly increase
Significant increase
WANG Qiang, et al.: The Three-North Shelterbelt Program and Dynamic Changes in Vegetation Cover
Increased vegetation cover from 1982 to 1991, 1992 to
2001, and 2002 to 2006 was 472 994, 451 032 and 435 196
km2, accounting for 11.62%, 11.09% and 10.70% of the
total study area. This was mainly the case in the central and
southern Greater Khingan Range, central Lesser Khingan
Range, Changbai Mountains in the northeast segment,
Yanshan Mountains, low mountains upland region of
western Liaoning Province, the Altai Mountains, Tien Shan
Mountains, Qilian Mountains (Shi et al. 2006; Guo et al.
2008), eastern of the northwest dessert, and the southern
Loess Plateau hilly-gully region. Vegetation here increased
primarily because the climate is moderate and precipitation
increased, and because vegetation construction projects
such as the TNSP and Green for Grain facilitate vegetation
recovery and growth (Zhu et al. 2004; Deng et al. 2006).
A decline in vegetation cover of 197 717, 270 192 and
182 690 km2 was found, accounting for 4.86%, 6.63% and
4.49% of the total study area. These areas were mainly
distributed on both sides of the Greater Khingan Range,
western Hulun Buir Plateau (Zhang et al. 2009), northern
Three River Plain, Horqin Sandy Land and southern
the northwest dessert and northern Loess Plateau hillygully region. The human population has grown extremely
fast in these areas, and combined with over grazing and
land reclamation, vegetation cover has declined. Further,
precipitation in these regions is low and its distribution is
uneven. In the last 25 years, precipitation has decreased at
the rate of 98.8 mm per 10 years, especially from 1992 to
N
0
Temperatur
500 1000
N
0
Correlation
coefficient
km
-0.81– -0.50
-0.50– -0.38
-0.38–0.38
0.38–0.50
0.50–0.75
km
2001 (rate of decline was 319.4 mm per 10 years). High
evaporation and high wind with a long duration have
aggravated drought in this region. As the climate continues
to warm, climate volatility in these regions will increase and
drought will be the main cause of vegetation cover decline.
4.4 Correlation between vegetation and climatic factors
The relationship between NDVI and temperature is
positive overall. A negative relationship between NDVI
and precipitation was found except for differences in
some regions (Fig. 6; Table 2). Regions with a negative
correlation between NDVI and temperature account for
approximately 17.74% of the total area and are mainly
distributed in the Junggar Basin, eastern Tarim Basin, Hexi
Corridor, western Qaidam Basin, Ulan Buh Sandy Land,
and northern Loess Plateau. Most of these regions are
located in arid or semiarid areas with high evaporation, high
temperatures, low precipitation, and irrigation agriculture,
leading to soil drying. On the contrary, the proportion of
the TNSP area with a positive correlation between NDVI
and temperature is 6.84%, and primarily distributed in the
Altai Mountains, Tien Shan Mountains, western Kunlun
Mountains, Qilian Mountains, eastern Qaidam Basin,
Qinling Mountains, the central Inner Mongolia Plateau,
southern Greater Khingan Range and Changbai Mountains.
These areas are similar in that they have a high latitude
and high altitude, suggesting that heat conditions are the
dominant natural factor influencing changes in vegetation.
Vegetation NDVI was negatively correlated with
precipitation across 10.60% of the total area, mainly
distributed in the Altun Mountains, western Alxa Plateau
and northern central Greater Khingan Range. Vegetation
NDVI was positively correlated with precipitation in
19.53% of the area, and this was found in the Tien Shan
Mountains, eastern Kunlun Mountain, eastern Qinghai
Province, Maowusu Sandy Land, Wulanchabu Plateau,
Xilingol Plateau, Horqin Sandy Land, northeast of Chifeng,
Songnen Plain, Liaohe River Plain and the southern part
of the Lesser Khingan Range. The different acquisition
ability of soil moisture results from the diversity of surface
Table 2 Correlation coefficient between average annual
NDVI, temperature and precipitation.
Precipitation
500 1000
57
Correlation
coefficient
-0.90– -0.50
-0.50– -0.38
-0.38–0.38
0.38–0.50
0.50–0.79
Fig. 6 Correlation between NDVI and climate factors
1982–2006.
Correlation coefficient distribution range
(%) Precipitation
(%)
Temperature
Significant negative –0.81– –0.50 6.44 –0.90– –0.50 0.88
correlation
Low negative
–0.50– –0.38 11.30 –0.50– –0.38 9.72
correlation
Weak correlation
–0.38–0.38 75.42 –0.38–0.38
69.87
Low positive
0.38–0.50
4.93
0.38–0.50
14.07
correlation
Significant positive 0.50–0.75
1.91
0.50–0.79
5.46
correlation
58
vegetation cover and leads to differences in precipitation
relevance (Li et al. 2005; Dai et al. 2010a, 2010b). In arid
and semiarid areas precipitation is a crucial constraint on
vegetation growth.
5 Conclusions
Regarding the temporal pattern, NDVI values across TNSP
region appear a slow upward trend during the last 25 years
and the growth rate is 0.007 per 10 years. The growth rate
of the wind-sand, northeast and north China plain areas
is most obvious, indicating that construction of the TNSP
is achieving and the trend of ecological degradation and
vegetation cover loss in most regions has been curbed. It is
necessary to make the TNSP sustainable and maintain effort
to reach the stated goals of this project.
There is a positive correlation between vegetation,
precipitation and temperature across the TNSP. For
example, for 17.74% of the region temperature was
negatively correlated.However, 6.84% of the region it was
positively correlated. Further, for 10.60% of the region
precipitation was negatively correlated and for 19.53%
of the region it was positively correlated. The area where
vegetation is closely associated with precipitation is
significantly larger than the area where temperature plays
the major role, and therefore precipitation is the key driver
of vegetation growth across the TNSP.
During the periods 1982–1991, 1992–2001 and 2002–
2006, vegetation cover area increased 472 994, 451 032
and 435 196 km2 respectively. This was mainly in central
and southern parts of the Greater Khingan Mountains, the
central Lesser Khingan Mountains, northeast section of the
Changbai Mountains, Yanshan Mountains, Western Liaoning
Hilly Region, Altai Mountain, Tien Shan Mountains, east
of the Qilian Mountains, east of the western desert area and
south of the hilly-gully area of the Loess Plateau. However,
in the same three periods, vegetation cover declined by
197 717, 270 192 and 182 690 km2, respectively, both sides
of the Greater Khingan Mountains, west of the Hulun Buir
Plateau, north of Sanjiang Plain, south of Horqin Sandy
Land, south of the western desert area and north of the
hilly-gully area of the Loess Plateau. Vegetation reduction
is affected by human activities as zones with obvious
reductions are places where the ecological environment is
weak and interference from human activity is intense. With
climate warming, climate fluctuations in these zones will
increase so the arid climate is another reason for vegetation
cover decline. These zones should be key management
areas into the future.
Acknowledgements
The authors thank the Environmental & Ecological Science Data
Center for West China and the Climatic Data Center, National
Meteorological Information Center, China Meteorological
Administration for data sets.
Journal of Resources and Ecology Vol.5 No.1, 2014
References
Adams J E, G F Arkin. 1997. A light interception method for measuring row
crop ground cover. Soil Science Society of America Journal, 41 (4): 789792.
Brown M E, D J Lary, A Vrieling, et al. 2008. Neural networks as a tool for
constructing continuous NDVI time series from AVHRR and MODIS.
International Journal of Remote Sensing, 29:7141-7158.
Carlson T N, D A Ripley. 1997. On the relation between NDVI, fractional
vegetation cover, and leaf area index. Remote Sensing of Environment,
62(3): 241-252.
Dai S P, Zhang B, Wang H J, et al. 2010a. Analysis on the spatio-temporal
variation of grassland cover using SPOT NDVI in Qilian Mountains.
Progress in Geography, 29(9):1075-1080. (in Chinese)
Dai S P, Zhang B, Wang H J. 2010b. Spatio-temporal change of vegetation
index NDVI in northwest China and its influencing factors. GeoInformation Science, 12(3):315-321. (in Chinese)
Defries R S, J R G Townshond. 1994. NDVI-derived land cover
classification at a global scale. International Journal of Remote Sensing,
15(17):3567-3586.
Deng C P, Guo N, Wang J M, et al. 2006. Vegetation cover variation in
northwest China during the past 22 years. Journal of Glaciology and
Geocryology, 28(5): 686-693. (in Chinese)
Guo Z X, Wang Z M, Song K S, et al. 2008. Changes of vegetation
coverage in northeast China from 1982 to 2003. Acta Botanica BorealiOccidentalia Sinica, 28(1): 155-163. (in Chinese)
Ichii K, Kawabata A, Yamaguchi Y. 2002. Global correlation analysis for
NDVI and climatic variables and NDVI trends: 1982-1990. International
Journal Remote Sensing, 23(18):3873-3878.
Jason E, G Roland. 2004. Discrimination between climate and humaninduced dryland degradation. Journal of Arid Environments, 57(4):535554.
Kaufman Y, D Tanre. 1992. Atmospherically resistant vegetation index
(ARVI) for EOS-MODIS. IEEE Transactions on Geoscience and Remote
Sensing, 30 (2): 261-270.
Lasaponara R. 2006. On the use of principal component analysis (PCA) for
evaluating interannual vegetation anomalies from SPOT/VEGETATION
NDVI temporal series. Ecological Modelling, 194(4):429-434.
Li J S. 1994. The state of development of the protective forest systems in
“Three Norths” and some countermeasures. World Forestry Research,
(1):64-89. (in Chinese)
Li Z, Yan F L, Fan X T. 2005. The variability of NDVI over north west
China and its relation to temperature and precipitation. Journal of Remote
Sensing, 9(3):308-313. (in Chinese)
Ma M G, Wang J, Wang X M. 2006. Advance in the inter-annual variability
of vegetation and its relation to climate based on remote sensing. Journal
of Remote Sensing, 10(3): 421-431. (in Chinese)
Pettorell N, J O Vik, A Mysterud, et al. 2005. Using the satellite-derived
NDVI to assess ecological responses to environmental change. Trends in
Ecology & Evolution, 20(9):503-510.
Shi Q D, Lv G H, Pan X L, et al. 2006. Characteristic analysis of vegetation
cover change in China western arid land. Science & Technology Review,
24(3):52-56. (in Chinese)
Song Y, Ma M G. 2008. Variation of AVHRR NDVI and its relationship
with climate in Chinese arid and cold regions. Journal of Remote Sensing,
12(3):499-505. (in Chinese)
Stow D, S Daeschner, A Hope, et al. 2003. Variability of the seasonally
integrated normalized difference vegetation index across the north slope
of Alaska in the1990s. International Journal of Remote Sensing, 24 (5):
1111-1117.
Stow D, A Petersen, A Hope, et al. 2007. Greenness trends of Arctic tundra
vegetation in the 1990s: Comparison of two NDVI datasets from NOAA
AVHRR systems. International Journal of Remote Sensing, 28(21): 48074822.
Tucker C J, J E Pinzon, M E Brown, et al. 2005. An extended AVHRR 8-km
NDVI dataset compatible with MODIS and SPOT vegetation NDVI data.
WANG Qiang, et al.: The Three-North Shelterbelt Program and Dynamic Changes in Vegetation Cover
International Journal of Remote Sensing, 26(20), 4485-4498.
Tucker C J, J R G Townshend. 1985. African land-cover classification using
satellite data. Science, 227(4685): 369-375.
Wang X H, Piao S L, P Ciaisb, et al. 2011. Spring temperature change and
its implication in the change of vegetation growth in North America from
1982 to 2006. PNAS, 108(4): 1240-1245.
59
Zhang Z X, Liu L M, Jia Y, et al. 2009. Climatic ecological adaptation of
shelter forests in Three-North Regions. Chinese Journal of Ecology,
28(9): 1696-1701. (in Chinese)
Zhu J Z, Zhou X C, Hu J Z. 2004. Thoughts and views about the three north
shelterbelt program. Journal of Natural Resources, 19(1): 79-85. (in
Chinese)
基于GIMMS AVHRR NDVI数据的三北防护林工程区植被覆盖动态变化与影响因子分析研究
王 强1，张 勃2，张志强1，张喜风3，戴声佩4
1 中国科学院资源环境科学信息中心/中国科学院国家科学图书馆兰州分馆，兰州 730000；
2 西北师范大学地理与环境科学学院，兰州 730070；
3 兰州大学西部环境教育部重点实验室旱区流域科学与水资源研究中心，兰州 730000；
4 中国热带农业科学院科技信息研究所，儋州 571737
摘 要：本文基于1982-2006年连续25年的GIMMS AVHRR NDVI植被覆盖指数，采用了最大化NDVI均值法、与气温及
降水变化的相关性和一元线性回归趋势分析法，对中国三北防护林工程区连续25年的植被覆盖时空变化特征进行了动态变化
研究。结果表明：（1）近25年来，研究区植被NDVI平均值总体呈缓慢上升趋势，增速为每10年0.007；（2）研究区植被和气
温、降水整体呈正相关关系，植被与降水正相关面积明显大于植被与气温正相关面积，说明降水是研究区植被生长的关键因
子；（3）1982-2006年，研究区植被覆盖增加的区域主要分布在大兴安岭中、南部，小兴安岭中部，长白山东北段，燕山，辽
西低山丘陵区，阿尔泰山，天山，祁连山东段，西北荒漠区东部和黄土高原丘陵沟壑区南部等；植被覆盖减少的区域主要是在
大兴安岭两侧，呼伦贝尔高原西部，三江平原北部，科尔沁沙地南端，西北荒漠区南部和黄土高原丘陵沟壑区北部等。
关键词：NDVI；三北防护林工程区；时空变化；相关分析；动态研究；中国北部