Spatial and temporal patterns of fuelwood collection in Wolong Nature Reserve: Implications for panda conservation
ABSTRACT Approximately 3 billion people, half of the World's population, are still using fuelwood in their daily lives. Fuelwood collection has been recognized as an important factor in habitat fragmentation and degradation and biodiversity loss, especially in developing countries. Understanding spatial and temporal patterns of fuelwood collection is fundamental to understanding human–environment interactions and designing effective conservation policies. Using Wolong Nature Reserve for giant pandas (Ailuropoda melanoleuca) in China as an example, we surveyed 200 rural households for the locations of their fuelwood collection sites in the past three decades (1970s, 1980s, and 1990s) and other ecological, economic, social, and demographic data. We found that fuelwood collection sites were becoming higher in elevation, more remote, and closer to highly suitable panda habitat from the 1970s to the 1990s. Consequently, fuelwood collectors were traveling longer distances to physically challenging areas, in our case, to areas of high-quality panda habitat. These spatial and temporal patterns of fuelwood collection suggest that future conservation policies for giant pandas, and other species worldwide, should also consider the needs of local communities.
- [Show abstract] [Hide abstract]
ABSTRACT: Regional ecological degradation induced by hydroelectric project construction (HPC) is of great concern in the field of landscape ecology research. Using GIS-based spatial analysis, we predicted and assessed the impacts of HPC on the ecological integrity of the Nuozhadu Nature Reserve (NNR). The results show that, after Nuozhadu HPC, the naturality of the NNR will be modified due to changes in the landscape composition such that larger areas covered by vegetation will be occupied by construction land and flooded by water areas. Meanwhile, landscape diversity will increase due to the additional landscape types of construction land and submerged areas, while landscape stability will decrease because of the splitting and contagion of the landscape after Nuozhadu HPC. The human disturbance index shows that the HPC will contribute to increasing the disturbance of the ecosystem. From buffer analyses, we conclude that the impacts of HPC will mainly occur in buffer zones over the distance of 800 m from the Lancang River in the NNR, and tend to be moderate in the 800–5,000 m buffer zone. Therefore, within the 800 m buffer zone, taking the naturality, diversity and stability of the ecosystem as well as anthropogenic interference as evaluation indicators, we calculated the ecological integrity index; the results indicate that the ecological integrity of the NNR will decrease by 7.6 % after project construction.Stochastic Environmental Research and Risk Assessment 27(7). · 1.96 Impact Factor
- Alzheimer's and Dementia 07/2011; 7(4). · 17.47 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Protected areas (PAs) often depend on landscapes surrounding them to maintain flows of organisms, water, nutrients, and energy. Park managers have little authority over the surrounding landscape although land use change and infrastructure development can have major impacts on the integrity of a PA. The need for scientifically-based regional-scale land use planning around protected areas is acute in human-dominated landscapes to balance conservation goals with livelihood needs for fuelwood, fodder, and other ecosystem services. As a first step, we propose the designation of a “zone of interaction” (ZOI) around PAs that encompasses hydrologic, ecological, and socioeconomic interactions between a PA and the surrounding landscape. We illustrate the concept by delineating the ZOI in three Indian PAs – Kanha, Ranthambore, and Nagarahole – using remote sensing, population census, and field data. The ZOI in Ranthambore is three times the size of the park and is largely defined by the socioeconomic interactions with surrounding villages. Ranthambore is located in headwaters and wildlife corridors are largely severed. In Nagarahole, the ZOI is more than seven times larger than the park and includes upstream watershed and elephant corridors. Kanha’s ZOI is approximately four times larger than the park and is mostly defined by contiguous surrounding forest. The three examples highlight the differing extents of ZOIs when applying equivalent criteria, even though all are located in densely-populated landscapes. Quantitative understanding of which activities (e.g. collection of forest products, grazing, road construction, tourism development) and which locations within the ZOI are most crucial to conservation goals will enable improved land use planning around PAs in human-dominated landscapes.Biological Conservation. 01/2010;
Landscape and Urban Planning 92 (2009) 1–9
Contents lists available at ScienceDirect
Landscape and Urban Planning
journal homepage: www.elsevier.com/locate/landurbplan
Spatial and temporal patterns of fuelwood collection in Wolong Nature Reserve:
Implications for panda conservation
Guangming Hea,∗, Xiaodong Chena, Scott Beaera, Manuel Colungab, Angela Mertigc, Li Ana,
Shiqiang Zhoud, Marc Lindermana, Zhiyun Ouyange, Stuart Gageb, ShuXin Lia, Jianguo Liua
aCenter for Systems Integration and Sustainability, Department of Fisheries and Wildlife, Michigan State University, East Lansing, MI 48824, USA
bComputational Ecology and Visualization Laboratory, Department of Entomology, Michigan State University, East Lansing, MI 48824, USA
cDepartment of Sociology & Anthropology, Middle Tennessee State University, Murfreesboro, TN 37132, USA
dChina’s Center for Giant Panda Research and Conservation, Wolong Nature Reserve, Wenchuan County, Sichuan Province, China
eDepartment of Systems Ecology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing, China
a r t i c l ei n f o
Received 16 March 2008
Received in revised form 13 January 2009
Accepted 22 January 2009
Available online 27 February 2009
Wolong Nature Reserve
a b s t r a c t
Approximately 3 billion people, half of the World’s population, are still using fuelwood in their daily lives.
and biodiversity loss, especially in developing countries. Understanding spatial and temporal patterns of
fuelwood collection is fundamental to understanding human–environment interactions and designing
effective conservation policies. Using Wolong Nature Reserve for giant pandas (Ailuropoda melanoleuca)
in China as an example, we surveyed 200 rural households for the locations of their fuelwood collection
sites in the past three decades (1970s, 1980s, and 1990s) and other ecological, economic, social, and
demographic data. We found that fuelwood collection sites were becoming higher in elevation, more
remote, and closer to highly suitable panda habitat from the 1970s to the 1990s. Consequently, fuelwood
collectors were traveling longer distances to physically challenging areas, in our case, to areas of high-
quality panda habitat. These spatial and temporal patterns of fuelwood collection suggest that future
conservation policies for giant pandas, and other species worldwide, should also consider the needs of
© 2009 Elsevier B.V. All rights reserved.
Approximately 3 billion people, half of the World’s popula-
tion, are still using fuelwood in their daily lives (Population Action
tant in ecological degradation worldwide, especially in developing
countries, where in many rural areas fuelwood is the sole or pri-
2001; An et al., 2002). Fuelwood collection through cutting down
trees can lead to fragmentation and degradation of wildlife habitat
(Liu et al., 2001), reduction of wildlife populations (Aigner et al.,
1998; Hall and Farrell, 2001), and loss of biodiversity (Rosenstock,
1998; Sagar and Singh, 2004).
Even many protected areas (e.g., nature reserves) in countries
such as China are not exempt from impacts of human activities
such as fuelwood collection (Liu et al., 2003b). In order to con-
∗Corresponding author at: Department of Fisheries and Wildlife, Center for Sys-
tems Integration and Sustainability, Michigan State University, 13 Natural Resources
Building, East Lansing, MI 48824, USA. Tel.: +1 517 432 5074; fax: +1 517 432 1699.
E-mail address: email@example.com (G. He).
serve its diverse natural resources for sustainable development,
China had established 2531 nature reserves by the end of 2007,
covering more than 15% of its territory (China.com.cn, 2007). How-
ever, many of these reserves are located in remote areas with
types of resource extraction are inevitable. Wolong Nature Reserve,
one of the largest reserves for protecting giant pandas (Ailuropoda
melanoleuca), is a good example. Fuelwood collection is common
inside the reserve and fluctuated around 7000–9400m3/year until
recently (Liu et al., 1999b). Bearer et al. (2008) showed that panda
activities in forests are reduced for several decades after timber
harvesting and fuelwood collection in this reserve. An et al. (2005)
found that fuelwood collection could lead up to a 1.23km2/year
loss of habitat, depending on different scenarios of socioeconomic
factors. Considering the distribution of bamboo and its periodic
flowering, Linderman et al. (2005) demonstrated that over the next
30 years fuelwood collection would result in the loss of up to 30%
of the habitat in the event of bamboo die-offs. If spatial arrange-
ment and configuration of habitat had been incorporated into such
an analysis, there would have been much higher impacts, as larger
areas have been fragmented from a landscape ecology perspec-
0169-2046/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
G. He et al. / Landscape and Urban Planning 92 (2009) 1–9
The knowledge of spatial and temporal distribution of fuelwood
collection at the landscape level is central to understanding the
impacts of fuelwood collection on forests (Franklin et al., 2000;
Pinzon et al., 2003) and panda habitat, including understory bam-
boo in the case of the giant panda (Reid et al., 1991; Taylor and Qin,
1993). This understanding can aid in evaluating habitat suitabil-
ity where human activities occur and providing insights to make
informed conservation decisions (e.g., identifying priority areas
for monitoring and altering human activities). Assessing the effec-
tiveness of previously implemented policies related to fuelwood
collection (Wolong Nature Reserve, 1998, 2000b) could also help
improve several on-going conservation programs in China (Loucks
et al., 2001; Zhu and Feng, 2002; Liu et al., 2008), including the Nat-
ural Forest Conservation Program (NFCP) and the Grain-to-Green
Program at Wolong. However, little has been done on gaining such
Therefore, this paper aims to: (1) characterize spatial and tem-
Reserve; (2) analyze the spatial and temporal trends of the impacts
of fuelwood collection on panda habitat; (3) discuss some policy
implications for panda conservation; (4) assess the general rele-
vance of the case study for fuelwood collection management in
2. Data and methods
2.1. Study area
Wolong Nature Reserve was designated in 1975 to help con-
serve the giant panda (MacKinnon and DeWulf, 1994). It is located
in Wenchuan County, Sichuan Province, southwestern China (lat-
itude: 30◦45?–31◦25?N, longitude 103◦52?–103◦24?E) (Fig. 1). The
Wolong Administration Bureau is responsible for both panda con-
servation and the well-being of local residents (Lü et al., 2003). It
reports directly to the Department of Forestry in Sichuan Province
Approximately 110 giant pandas, representing about 10% of
the total wild population, inhabit the reserve (China’s Ministry of
Forestry and WWF, 1989; Zhang et al., 1997). The vegetation of the
reserve includes evergreen broadleaf, deciduous, and sub-alpine
coniferous forests, and alpine meadows within an elevation range
of 1250–6525m above sea level (Schaller et al., 1985). Forests cov-
ered 36.3% of the reserve in 2001 (Vina et al., 2007), less than 1% of
the land is for agricultural use, and the remainder is shrubs, mead-
ows, permanent snow, exposed rocks, roads, buildings, and water
for cover and shelter and use the understory bamboo (mainly
Bashania fangiana and Fargesia robusta) as staple food (Schaller
et al., 1985). They prefer slopes less than 30◦, elevations between
1500 and 3250m, and interior, old-growth forests (Schaller et al.,
1985; Bearer et al., 2008). Panda habitat is determined by biotic
features (e.g., forests), abiotic features (elevations and slopes),
and human activities (e.g., roads). The widely used panda habi-
tat classification scheme suggested by Liu et al. (1999b) includes
four categories: highly suitable, moderately suitable, marginally
suitable, and unsuitable. Highly and moderately suitable habitats
have conifer forests and the two main bamboo species, whereas
marginal habitats have evergreen broadleaf forests and other bam-
boo species (Bashania fangiana and Fargesia robusta). Unsuitable
habitat is defined as having no forest cover and no bamboo, an ele-
vation >3750m, and a slope>45◦(see Liu et al., 1999b, for details
on habitat classification). Due to the lack of data on bamboo distri-
bution for the entire reserve over time, forest cover, elevation, and
slope are often used for panda habitat suitability analysis (Liu et al.,
2001; Vina et al., 2007, 2008). One of such efforts (Vina et al., 2007)
shows that approximately 4.4% of the reserve was marginally suit-
able, 25.6% was moderately suitable, and 5.8% was highly suitable,
and the rest was unsuitable for pandas in 2001.
Timber harvesting, poaching, and agriculture were the main
threats to panda habitat and conservation. Recently, fuelwood col-
lection by local residents has emerged as an important threat. From
1975 to 1998, the human population and the number of house-
holds in the reserve increased by 69 and 124%, respectively, while
fuelwood consumption in the reserve doubled (Liu et al., 1999a,b).
Traditionally, fuelwood is used for cooking food for humans and
fodder for livestock, and for heating houses in winter (An et al.,
2002). In 2000, there were more than 4400 local rural residents
in 970 households in two townships (Wolong and Gengda, see
Fig. 1 for their locations). Most human settlements were in the
bottomland or on the relatively flat slopes of several valleys in the
villages, and townships. Wolong Township has 9 groups in three
villages, and in Gengda Township has 17 groups in three villages
(Wolong Nature Reserve, 2000a). The annual amount of fuelwood
consumed by each household ranged from 8 to 30m3, depending
nomic conditions (An et al., 2001). Moreover, burgeoning tourism
Fig. 1. The location and elevations of Wolong Nature Reserve in China. The locations of Wolong and Gengda townships are also indicated on the map.
G. He et al. / Landscape and Urban Planning 92 (2009) 1–9
development now may be contributing to greater fuelwood col-
lection by local residents, who can increase their cash income by
selling local products, such as bacon, which requires fuelwood to
cook pig fodder and smoke-dry the pork (He et al., 2008).
The household is the basic unit of fuelwood consumption (Liu et
al., 2003a), but fuelwood collection is often accomplished in win-
ter by groups of 10–20 adult males each from several households to
increase efficiency and minimize risks in the topographically chal-
lenging high mountains (Liu et al., 1999b). People usually walk or
drive to the foothills close to the collection sites and climb up to
those sites. They cut down trees with axes, then carry and slide logs
vehicles, and transported home. Later, fuelwood logs are usually
split into small pieces and piled near houses.
Local residents usually cut down large trees, partially remov-
ing forest canopy (An et al., 2001). Oak (Cyclobalanopsis Oerst.),
Carr.), larch (Larix Mill.), and pine (Pinus L.) are among the preferred
tree species for fuelwood. Fuelwood collection changes the species
composition in the overstory, thus stimulates denser understory
bamboo stands with lower moisture content, and, consequently,
discourages pandas from using the affected areas (Reid et al., 1991).
Although selective cuts of a few trees for fuelwood at one time by
one household may affect only a small area of habitat, collection of
fuelwood by multiple households over a long time (e.g., decades)
can eventually lead to the loss of a large amount of habitat (Bearer
et al., 2008). Fuelwood collection started in the forests close to
human settlements along the main road through the reserve but
has expanded to other areas and caused a significant reduction in
the quality and quantity of panda habitat (Liu et al., 1999a).
2.2. Data collection
We used historical information on fuelwood collection over the
past three decades (1970s, 1980s, and 1990s) as well as other eco-
logical, economic, social, and demographic data to examine spatial
and temporal patterns of fuelwood collection. One type of data is
at the household level, such as household location (An et al., 2005),
population (Wolong Nature Reserve, 2000a), fuelwood collection
sites of the last three decades, attitudes toward fuelwood poli-
cies, and household economy of 2001. Another type of data is at
the reserve level, which consists of digital elevation model (DEM),
interpolated from digitized 100m contours with a vertical accu-
racy of <50m root-mean-square error (RMSE)) (An et al., 2005),
habitat suitability index (HSI) maps, including four categories of
habitat – highly suitable, moderately suitable, marginally suitable,
and unsuitable – of 1974 and 1997 (Liu et al., 2001), and road net-
work and construction records.
2.2.1. Household survey
Intensive face-to-face household interviews were conducted in
the summer of 2002. Two hundred households (about one-fifth
of the total) were sampled based on a random sampling method
stratified by villages and groups using the population census data
of 2000. Of the 200 households interviewed, 12 established in the
1980s did not collect fuelwood in the 1970s, 8 did not collect in the
1990s, and 1 had not collected since the 1980s. Because of this, the
degrees of freedom for tests of different time combinations may be
and 388 for the comparison of the 1980s versus the 1990s). We
recorded the most frequently visited fuelwood site for a household
a scale of 1:50,000. The site locations for all the surveyed house-
holds were then digitized and geometrically corrected to form a
geographic information system (GIS) layer in ArcGIS.
Eight activities not allowed in fuelwood collection.
Do not cut newly planted seedlings for fuelwood
Do not cut young trees for fuelwood
Do not cut trees in the previously harvested areas for fuelwood
Do not cut trees in the erosion-prone areas for fuelwood
Do not cut trees of rare and precious species for fuelwood
Do not cut riparian trees for fuelwood
Do not cut trees in research zones for fuelwood
Do not light a fire in forests
Source: Translated from government documents (in Chinese) (Wolong Nature
We also simultaneously investigated the attitudes of the house-
hold heads, who usually know more about fuelwood collection
activities than other household members, toward two fuelwood-
related regulations to examine whether these policies affected the
spatial and temporal distribution of fuelwood collection. The first
regulation designated the allowable amount, time, and location of
do eight specific activities during collection (see Table 1). The reg-
ulations were released in 1984 (Wolong Nature Reserve, 2000b),
about halfway through our study period. We also asked household
heads whether they: (1) knew about the regulations, (2) if yes, how
long they had known about them, and (3) what effects the two reg-
ulations had on their fuelwood collection behaviors (e.g., amount
collected, time spent, and species of tree collected).
Besides the attitudes toward these two regulations, we also sur-
veyed the household economic status of 2001 to see if relatively
richer households were different from the poorer ones in fuelwood
collection activities of the 1990s (economic data of households in
the 1970s and 1980s were not available). We indirectly gathered
data on economic status during interviews by determining the sum
of annual household expenditures for all items, such as food, edu-
cation, energy consumption, farming, and medical services. When
collecting this retrospective information, we incorporated a widely
used method in social sciences, the life history calendar, to improve
the respondents’ recall accuracy (Caspi et al., 1996; Axinn et al.,
1999). We often asked the interviewees some indirect questions
on life events more readily remembered, such as births, marriages,
deaths, household separations, to help them recall less salient facts
related to our research interests, for example, the location of fuel-
wood collection and the year they knew a regulation was released.
We recorded global positioning system (GPS) measurements
and current status of fuelwood sites in the forest stands (num-
ber of stumps left, stump diameters, and tree species) to verify
whether sites recorded from interviews were actually used before.
We visited and surveyed fuelwood sites of 15 randomly selected
participants in fuelwood collection), we located those sites around
a heavily collected spot. Old stumps (from the 1970s) were iden-
tified by new seedlings developed from them and seedlings also
helped us to recognize tree species. We found that on average there
were 56.5 stumps left with a mean diameter of 16.2cm in each site.
The average number of tree species in these sites was 9.7. In total,
we identified 63 species of trees collected, of which seven were
in fuelwood piles during our household interviews. Therefore, we
are confident that the sites were used for fuelwood collection.
2.2.2. Road network
Records of transportation system development in the reserve
were kept in the Department of Transportation in Wolong (Wolong
Nature Reserve, 2005). Road quality, construction/improvement
G. He et al. / Landscape and Urban Planning 92 (2009) 1–9
time, and construction goals were well documented and road net-
works in different decades were mapped. Household surveys and
all other non-spatial attribute data were managed in a Microsoft
Access 2000 database. Spatial information (e.g., fuelwood sites,
road networks, and the DEM) was warehoused in an ArcGIS 8.2
database for further analyses.
2.3. Data analysis
2.3.1. Fuelwood collection sites
To characterize the topography of fuelwood collection sites,
average elevation, aspect, and slope in 30m×30m area (0.09ha)
around the sites were derived from the DEM. We chose this size as
a conservative estimation of annual collection area for a household
consuming a minimum amount of fuelwood (8m3) (An et al., 2001)
and collecting a relatively large percent (around 60%, field observa-
tions, 2002) of trees in a mixed evergreen and coniferous forest
stand (most frequently visited, with an average stock of approxi-
mately 120m3/ha) (Bearer, 2005). This selection also matches the
resolution (30m×30m) of available data (DEM and HSI map).
We used the point density function from Spatial Analyst Tools
in ArcGIS 8.2 to generate density maps of fuelwood collection sites.
The densities were arbitrarily categorized into three groups: 0–5,
5–10, and 10+ points per km2. Then we identified newly emerging
heavily used areas, usually centered by points with densities of 10+
points per km2as fuelwood collection hotspots by comparing the
maps in different decades.
2.3.2. Distances between fuelwood collection sites and household
Because the household is a basic unit of many human activ-
ities in the reserve (Liu et al., 2003a), their characteristics (e.g.,
location, economic status, labor) could be important in fuelwood
site selection. To understand the temporal and spatial relationships
between fuelwood sites and household locations, three distances
were measured using ArcGIS. ED is defined as the Euclidean dis-
tance between a household and its fuelwood collection site. TD
Euclidean distance from a fuelwood collection site to the location
of the nearest household, which may not be the focal household.
ED and TD were designed to describe the relationships between
individual households and their fuelwood collection sites at the
holds and fuelwood collection locations at the reserve landscape
We used information on household expenditures of 2001 as a
proxy of economic status of the 1990s. We categorized households
8.25RMB=US$ 1 in 2001), medium (>5000 and ≤10,000RMB), and
high (>10,000RMB). Annual per capita expenditures of 2001 were
also classified into three categories: low (≤1400RMB), medium
(>1400 and ≤2800RMB), and high (>2800RMB). The number of
laborers in a household in 2000, derived from the 2000 popula-
tion census data, was used to group the households to test whether
it affected households in terms of fuelwood collection in the 1990s
(data on the laborers in the 1970s and 1980s are not available). We
distinguished households with two or fewer laborers from those
with three or more laborers.
2.3.3. Relationships between fuelwood collection and panda
To examine the spatial and temporal trends of the impacts of
fuelwood collection on panda habitat, we measured two variables.
First, we calculated the distance (HD) from a fuelwood site to
the nearest pixel of each type of habitat on the HSI map of 1974.
Second, we derived the percentages of households collecting fuel-
wood in a certain type of habitat in a given decade by overlaying
habitat at the reserve landscape level rather than ED or TD at the
We used t-tests and analysis of variance (ANOVA) to test for dif-
ferences of elevation, slope, ED, TD, and ND of the three decades. As
an exception, the Mardia–Watson–Wheeler test was applied to test
for the differences of aspects of fuelwood sites in the three differ-
1981). In addition, we calculated and reported descriptive statistics
on attitudes of local residents toward fuelwood policies.
3.1. Fuelwood collection sites
Our results on the temporal changes of three topographic char-
acteristics (elevation, slope, and aspect) of fuelwood sites suggest
an average elevation>2000m and mean slope>30◦(Fig. 2). The
one-way ANOVA test rejects the null hypothesis that the means
of the three groups of elevations were equal (F=47.78, p<0.0001).
The t-tests (1970s versus 1980s [t=6.57, p<0.0001] and 1980s ver-
increased over time (on average, 100–150m per decade) (Fig. 2a).
The slopes of the three decades were not significantly different
(F=2.73, p=0.066) while they increased only from the 1970s to the
1980s at the significance level of 0.05 (t=2.33, p=0.016) (Fig. 2b).
The Mardia–Watson–Wheeler test shows that there were no sig-
nificant differences among the aspects of fuelwood sites during the
three decades studied (W=5.92, p=0.21) (Fig. 2c).
aspect (c) with 1 standard deviation. Two stars indicate that the value to their left
was significantly smaller than the one to their right at the significance level of 0.05.
G. He et al. / Landscape and Urban Planning 92 (2009) 1–9
Fig. 3. Fuelwood collection site density maps of the 1970s (a), 1980s (b), and 1990s (c), and emerging fuelwood hotspots in the 1980s (b), and 1990s (c). The density unit is
points per square kilometer.
Density maps of fuelwood sites used during the 1970s, 1980s,
and 1990s clearly indicate clustering of fuelwood collection across
the reserve landscape (Fig. 3). Several fuelwood hotspots had
emerged over time in the reserve (b and c in Fig. 3). We iden-
tified Area 1 as a hotspot in the 1980s by comparing the maps
in the 1970s and 1980s (Fig. 3b). Similarly, two hotspots, Area 2
and Area 3, attracted many households for fuelwood in the 1990s
(Fig. 3c). Furthermore, linking road construction information with
these hotspots shows that road development might have con-
tributed to their emergence. Roads near Area 1 were constructed
from a trail to improve local access to markets with support from
in the early-1980s (Wolong Nature Reserve, 2005), but extended so
far into the forests that they clearly facilitated fuelwood collection
in Area 1. Roads near Area 2 were originally developed for mining
G. He et al. / Landscape and Urban Planning 92 (2009) 1–9
Three types of distances between fuelwood sites of the 1990s and locations of households grouped by the total household expenditure of 2001 (A), household per capita expenditure of 2001 (B), and number of laborers in each
household in 2000 (C).
A (total household expenditure, in RMBa)
B (household per capita expenditure)
C (number of laborers)
ED (Euclidean distances between fuelwood
sites and households, m)
TD (travel distances between fuelwood sites
and households, m)
ND (Euclidean distances between fuelwood
sites and their nearest households, m)
Note: For each type of distance measurement, in the first row, the number outside the parentheses is the mean of the distances in the group and the number inside is the standard deviation; the second row shows the results of
one-way ANOVA tests of the three groups by the factor A, B, or C. The group factors A, B, and C are also shown in the first row of the table, and the group definitions for each factor are given as divided columns below the factors.
a1RMB=approx. 15 cents U.S. currency.
Fig. 4. Temporal dynamics of distances (ED defined as the Euclidean distance
between a household and its fuelwood collection site, TD defined as the length of
networks, and ND defined as the Euclidean distance from a fuelwood collection site
locations with 1 standard error of mean. Two stars indicate that the measurement
to their left was significantly smaller than the one to their right at the significance
level of 0.05, while one star indicates that at the significance level of 0.10.
1990s (Wolong Nature Reserve, 2005). The quality of the main road
across the reserve, from the east side to the lower southwest corner
and constructed for commercial timber logging in the 1960s before
1990s (it was paved, which was rare in western China during that
time). These roads fragmented the reserve, and facilitated house-
hold fuelwood collection considerably. People living far away from
the road could easily travel to Area 3, which had high forest stock
and was rarely used as a fuelwood site before the mid-1990s.
3.2. Distances between fuelwood collection and household
Besides climbing higher into the mountains, as mentioned
above, local residents also traveled greater distances (TD) to fuel-
wood sites, because sites became increasingly farther away from
households (ED) over time (Fig. 4). On average, local households
had to reach out an extra 20–50m farther (ED) or travel 50–80m
more (TD) yearly to find good fuelwood sites. Although the dis-
mountainous areas, which meant greater hardship and more time
required for fuelwood collection. Statistical evidence concerning
the distance between a fuelwood site and its nearest household
(ND) shows that, at the reserve level, fuelwood collection was
expanding gradually farther away from residential areas (Fig. 4).
From further analyses of spatial characteristics ED, TD, and ND,
grouped by economic status of 2001 (annual household and per
capita expenditures) and number of laborers in 2000, we found
no significant differences between rich and poor households, or
households with more or fewer laborers in terms of fuelwood
site selection in the 1990s (Table 2). Presumably, this is because
Fig. 5. Percentages of fuelwood sites of three decades (1970s, 1980s, and 1990s)
falling in four types of habitat (highly suitable, moderately suitable, marginally
suitable, and unsuitable) in 1974.
G. He et al. / Landscape and Urban Planning 92 (2009) 1–9
fuelwood collection was accomplished by groups of households
they were poor or rich, with more or fewer laborers.
3.3. Relationships between fuelwood collection and panda habitat
The percentage of fuelwood sites in highly suitable panda habi-
(Fig. 5). Most of fuelwood collection (varying between 68 and 78%)
occurred in areas of highly suitable and moderately suitable panda
habitat and much smaller percentages (ranging from 22 to 32%) in
areas of unsuitable habitat (Fig. 5). Surprisingly, no fuelwood col-
lection appeared in any areas of marginally suitable habitat. The
reason might be because marginally suitable habitat is character-
perspective of the distance between a fuelwood site and its nearest
habitat type, we found that local residents collectively went closer
to highly suitable habitat (144, 132, and 105m for the 1970s, 1980s,
and 1990s, respectively) and farther from unsuitable habitat (90,
127, and 145m for the 1970s, 1980s, and 1990s, respectively) to
find good sites for fuelwood over time.
To further demonstrate the spatial trends of fuelwood collec-
tion impacts on panda habitat, we calculated the distance between
a household and its nearest habitat pixel of a given type. The loca-
tions of households were relatively stable. Our results show that
for highly suitable habitat, the average distance between house-
holds and the 1974 habitat (153.43±131.59m) was significantly
shorter than the one between households and the 1997 habitat
(330.86±264.58) (t=12.80, p<0.001). Similarly, for moderately
suitable habitat, the mean distance corresponding to the 1974
Because the data for panda habitat in the 1980s were not available,
we only analyzed the habitat change of the fuelwood sites from the
1970s to the 1990s by overlaying the fuelwood sites on the habitat
maps in both 1974 and 1997. This analysis shows that more than
50% of fuelwood sites with highly suitable habitat in the 1970s had
that were moderately suitable in the 1970s had become unsuit-
able and the rest were unchanged. In total, the quality of 18.8% of
marize, the continual collection of fuelwood in the 1970s, 1980s,
and 1990s had degraded habitat quality and affected the spatial
distribution of panda habitat; primarily, suitable habitat had been
pushed farther away from households.
3.4. Knowledge about and attitudes toward fuelwood-related
Dissemination of fuelwood-related policies to local residents
within the reserve was effective. Among the 200 household heads
interviewed, 134 (67%) said they had known about at least one reg-
ulation. Among these 134 respondents, 84 had known about the
regulations for less than 10 years, while the rest had known about
them longer than 10 years. However, as to the effects of the poli-
cies on fuelwood collection, only 42% of those interviewed (84 of
200) or 63% of those who knew the regulations (84 of 134) said
that their fuelwood collection activities had been affected. Even
of them (56%) believed that less fuelwood was harvested because
of the existence of regulations, while for others regulations had no
impact on the amount of fuelwood collected. Regarding the effect
who followed the regulations thought that they spent less time
while 31% spent more time and 26% saw no change. The major
changes brought about by the policies were that the local residents
went to more remote sites (77% of households who followed the
ity of the resource in places nearby might have contributed to the
spatial trend of fuelwood collection activities moving farther away
from households over time in the reserve.
4. Discussion and implications
Our results reveal that local residents in Wolong Nature Reserve
had to select gradually more distant sites at increasingly higher
elevations to collect fuelwood during the last three decades of the
20th century, and most of tree fellings occurred in areas of highly
suitable and moderately suitable panda habitat rather than areas
of marginally suitable and unsuitable habitat. Fuelwood collec-
tion occurred frequently in areas close to households, while some
new hotspots had emerged due to local road expansion. Many
households were aware of the fuelwood collection regulations and
understood their importance to panda conservation, but many of
them did not comply strictly with them. As good forests receded
from households at the reserve level, local residents experienced a
gradually increasing hardship in their fuelwood collection, which
exacerbated an already-existing conflict between people and pan-
das. The implications of our results for policy are discussed below.
4.1. Road construction
Our study suggests that road systems facilitated fuelwood col-
lection in suitable panda habitat. Therefore, road extension should
be carefully planned, especially in conservation-oriented areas,
such as Wolong Nature Reserve. Investment should be focused on
improving the roads that may facilitate the access of local residents
to markets to sell their agricultural products (e.g., cash vegetables
like cabbage). Roads should not be expanded into areas without
residents or into areas from which residents have been encouraged
to move. Also, roads originally designed for mining or other non-
illegal use for resource extraction.
4.2. Energy-related policies
Over time, local residents have experienced increasing difficul-
ties in collecting fuelwood, involving tedious trips to steep and
remote mountains. Energy is a crucial issue for both human wel-
and serious, as fuelwood collection is no longer officially allowed.
In 2001, the Natural Forest Conservation Program was initiated in
the reserve. Each rural household was assigned one or more for-
est parcels to monitor for illegal harvesting and given a yearly
subsidy (available until 2010) equal to one-quarter or more of
household annual income. This direct income gained from NFCP
along with serious law enforcement has prompted most house-
holds to reduce fuelwood consumption since 2001 by switching
to electricity (Wolong Nature Reserve, 2005). The average electric-
ity consumption per household increased from 800kWh in 2000 to
from 1.4 to 0.3m3during the same period (Wolong Nature Reserve,
2005). A new, so-called ecohydropower plant began operation in
October of 2002, but the price of electricity, a widely proposed sub-
stitute (An et al., 2002; Wolong Nature Reserve, 2005), is still high
relative to household income (about 90% of the local rural residents
interviewed complained about it). The effect of lowering the price
of electricity on panda habitat can be significant. For example, a
simulation study indicated that a reduction in electricity price by
G. He et al. / Landscape and Urban Planning 92 (2009) 1–9
0.05RMB could save approximately 15km2of panda habitat over a
period of 30 years (An et al., 2006).
With substantial income losses when the NFCP ends, the house-
holds will likely resile and cut down trees for fuelwood again.
Economically affordable, socially acceptable, ecologically sound,
and sustainable long-term alternatives are needed. An on-going
Grain-to-Green Program applies a national policy that focuses on
restoring ecological functions as well as encouraging economic
development in steep cropland; the program requires at least 80%
of trees planted for ecological restoration, and at most, 20% of trees
for economic purposes (Zhu and Feng, 2002; Ye et al., 2003). How-
ever, it may not effectively address the local residents’ main need:
fuelwood in Wolong Nature Reserve. As suggested by some other
scholars for other regions (Madeschin, 1999; Zhang et al., 2000;
Kohlin and Parks, 2001; Richardson et al., 2002), we also recom-
mend that most of steep cropland returned to vegetation should
be designated specifically for growing fuelwood forests, providing
both energy resources and ecological functions in the reserve. By
continuing to use fuelwood, local residents (of three minority eth-
nic groups: Tibetan, Qiang, and Hui) may preserve their traditional
lifestyles and cultures (Ngugi, 1988; Mahiri and Howorth, 2001).
4.3. Panda habitat conservation
Considering the reality that fuelwood collection is unavoidable,
the impacts on panda habitat could be mitigated if more fuelwood
collection were to occur in areas with panda habitat of relatively
lower quality, where little fuelwood was collected in the past. It is
evident that fuelwood policies did not change the local residents’
fuelwood collection behaviors. There were three major reasons for
the apparent failure of these regulations. First, law enforcement
was weak: fuelwood collection occurred in topographically dif-
ficult and relatively large areas, but the Wolong Administration
Bureau had limited staff for resource monitoring and protection
(Wolong Nature Reserve, 2005). Second, the tragedy of the com-
mons (Hardin, 1968) is applicable in Wolong: collecting fuelwood
based on the current Chinese land system, local residents have only
usufruct rights; therefore, there is little incentive for them to grow
trees for fuelwood, use energy-efficient stoves, or reduce fuelwood
not well addressed (Sharma, 1990): many potential alternatives
(e.g., biogas and wind/sun power) were unavailable, were expen-
sive (e.g., coal, because of transportation), or had other negative
environmental effects (e.g., greenhouse gases from burning char-
coal and coal). Although the eight hydropower plants in the reserve
had a total capacity of 33,960kW in 2003 (Wolong Nature Reserve,
tricity was sold to cities. At the average price of 0.13RMB/kWh in
of households switching to electricity (An et al., 2002).
fuelwood collection in the areas of highly suitable and moderately
suitable habitat. Currently, the farther an NFCP parcel is from the
receives. To improve the efficacy of the NFCP’s investment, house-
hold should receive higher subsidies if their parcels are in better
shape, i.e., the habitat quality is better.
Meanwhile, we suggest that households be officially allowed to
collect dead tree branches and shrubs for fuelwood from nearby
NFCP parcels, with a maximum limit. This compromise should be
allowed only for a short period, because in the long run it may also
et al., 1997; Aigner et al., 1998; Rosenstock, 1998; Kumar and
Shahabuddin, 2005). This proposal is most likely feasible because
of the successful implementation of NFCP: forest parcels were well
monitored and few were illegally harvested. This will also give
to grow for fuelwood.
Understanding spatial and temporal patterns of fuelwood col-
lection will not only help researchers scale up their evaluations
of the impacts of fuelwood collection on panda habitats from the
damental knowledge for reserve managers to make informed and
effective decisions, and evaluate and adjust policies. For example,
types of habitat quality affected by fuelwood collection. Further-
more, the results obtained from and the methods used in this study
may be useful for similar efforts in analyzing temporal and spatial
patterns of fuelwood collection or other human activities in other
parts of the world.
We appreciate the help in the process of data entry and anal-
ysis from Gang Tang and Hongxia Lü. Our thanks also go to the
Wolong Nature Reserve Administration for the logistic support of
our fieldwork and Xiaoping Zhou, Mingchong Liu, Lun Wang, Jian
Yang, Hemin Zhang, and Yingchun Tan for their assistance of field
work in the reserve. We also would like to thank three anonymous
reviewers for their constructive comments and suggestions and
Joanna Broderick for her editorial assistance. In addition, we grate-
fully acknowledge the financial support from the National Science
Foundation (CAREER Award and Biocomplexity in the Environment
Grant), and National Institutes of Health (National Institute of Child
Health and Human Development, R01 HD39789).
Aigner, P.A., Block, W.M., Morrison, M.L., 1998. Effect of firewood harvesting on birds
in a California oak-pine woodland. J. Wildlife Manage. 62 (2), 485–496.
An, L., Liu, J.G., Ouyang, Z.Y., Linderman, M., Zhou, S.Q., Zhang, H.M., 2001. Simulating
for giant panda habitats. Ecol. Model. 140 (1–2), 31–49.
An, L., Lupi, F., Liu, J.G., Linderman, M.A., Huang, J.Y., 2002. Modeling the choice
to switch from fuelwood to electricity: implications for giant panda habitat
conservation. Ecol. Econ. 42 (3), 445–457.
An, L., Linderman, M., Qi, J.G., Shortridge, A., Liu, J.G., 2005. Exploring complexity
in a human–environment system: an agent-based spatial model for multidisci-
plinary and multiscale integration. Ann. Assoc. Am. Geogr. 95 (1), 54–79.
An, L., He, G.M., Liang, Z., Liu, J.G., 2006. Impacts of demographic and socioeconomic
cations. Soc. Sci. Res. 28 (3), 243–264.
Batschelet, E., 1981. Circular statistics in biology. In: Sibson, R., Cohen, J.E. (Eds.),
Mathematics in Biology. Academic Press, Inc., London, p. 371.
Bearer, S.L., 2005. The Effects of Forest Harvesting on Giant Panda Habitat Use in
Wolong Nature Reserve, China. Department of Fisheries and Wildlife, Michigan
State University, East Lansing, MI, p. 173.
Bearer, S., Linderman, M., Huang, J.Y., An, L., He, G.M., Liu, J.Q., 2008. Effects of fuel-
141 (2), 385–393.
Bengtsson, J., Persson, T., Lundkvist, H., 1997. Long-term effects of logging residue
Caspi, A., Moffitt, T.E., Thornton, A., Freedman, D., Amell, J.W., Harrington, H.,
Smeijers, J., Silva, P.A., 1996. The life history calendar: a research and clini-
cal assessment method for collecting retrospective event-history data. Int. J.
Method. Psych. 6 (2), 101–114.
lished in China (before the end of 2007)), http://www.china.com.cn/txt/2008-
05/23/content 15419786.htm, Last Access: 09-13-2008.
(2531 Nature Reserves Estab-
and management plan for giant pandas and their habitat), Beijing, China (in
Chomitz, K.M., Griffiths, C., 2001. An economic analysis and simulation of woodfuel
management in the Sahel. Environ. Resour. Econ. 19 (3), 285–304.
G. He et al. / Landscape and Urban Planning 92 (2009) 1–9
Franklin, S.E., Moskal, L.M., Lavigne, M.B., Pugh, K., 2000. Interpretation and clas-
sification of partially harvested forest stands in the Fundy Model Forest using
Hall, T.E., Farrell, T.A., 2001. Fuelwood depletion at wilderness campsites: extent and
potential ecological significance. Environ. Conserv. 28 (3), 241–247.
Hardin, G., 1968. The tragedy of the commons. Science 162 (3859), 1243–1248.
He, G.M., Chen, X.D., Liu, W., Bearer, S., Zhou, S.Q., Cheng, L.Y., Zhang, H.M., Ouyang,
Z.Y., Liu, J.G., 2008. Distribution of economic benefits from ecotourism: a case
study of Wolong Nature Reserve for giant pandas in China. Environ. Manage. 42
tion and fuelwood collection in south Asia. Land Econ. 77 (2), 206–218.
Kumar, R., Shahabuddin, G., 2005. Effects of biomass extraction on vegetation struc-
Conserv. 32 (3), 248–259.
Linderman, M.A., An, L., Bearer, S., He, G.M., Ouyang, Z.Y., Liu, J.G., 2005. Modeling
the spatio-temporal dynamics and interactions of households, landscapes, and
giant panda habitat. Ecol. Model. 183 (1), 47–65.
Liu, J.G., Ouyang, Z.Y., Tan, Y.C., Yang, J., Zhang, H.M., 1999a. Changes in human pop-
ulation structure: implications for biodiversity conservation. Popul. Environ. 21
Liu, J.G., Ouyang, Z.Y., Taylor, W.W., Groop, R., Tan, Y.C., Zhang, H.M., 1999b. A frame-
work for evaluating the effects of human factors on wildlife habitats: the case of
giant pandas. Conserv. Biol. 13 (6), 1360–1370.
Liu, J.G., Linderman, M., Ouyang, Z.Y., An, L., Yang, J., Zhang, H.M., 2001. Ecological
degradation in protected areas: the case of Wolong Nature Reserve for giant
pandas. Science 292, 98–101.
Liu, J.G., Daily, G.C., Ehrlich, P.R., Luck, G.W., 2003a. Effects of household dynamics
on resource consumption and biodiversity. Nature 421 (6922), 530–533.
Liu, J.G., Ouyang, Z.Y., Pimm, S.L., Raven, P.H., Wang, X.K., Miao, H., Han, N.Y., 2003b.
Protecting China’s biodiversity. Science 300 (5623), 1240–1241.
Liu, J.G., Dietz, T., Carpenter, S.R., Alberti, M., Folke, C., Moran, E., Pell, A.N., Deadman,
P., Kratz, T., Lubchenco, J., Ostrom, E., Ouyang, Z., Provencher, W., Redman, C.L.,
Schneider, S.H., Taylor, W.W., 2007. Complexity of coupled human and natural
systems. Science 317 (5844), 1513–1516.
effects of China’s policies for ecosystem services. Proc. Natl. Acad. Sci. U.S.A. 105
Loucks, C.J., Lü, Z., Dinerstein, E., Wang, H., Olson, D.M., Zhu, C., Wang, D., 2001. Giant
pandas in a changing landscape. Science 294, 1465.
Lü, Y., Chen, L., Fu, B., Liu, S., 2003. A framework for evaluating the effectiveness of
protected areas: the case of Wolong Biosphere Reserve. Landscape Urban Plan.
63 (4), 213–223.
In: Miller, R.I. (Ed.), Mapping the Diversity of Nature. Chapman & Hall, London,
England, pp. 142–172.
Mahiri, I., Howorth, C., 2001. Twenty years of resolving the irresolvable: approaches
to the fuelwood problem in Kenya. Land. Degrad. Dev. 12 (3), 205–215.
Ngugi, A.W., 1988. Cultural-aspects of fuelwood shortage in the Kenyan highlands.
J. Biogeogr. 15 (1), 165–170.
Pinzon, Z.S., Ewel, K.C., Putz, F.E., 2003. Gap formation and forest regeneration in a
Micronesian mangrove forest. J. Trop. Ecol. 19, 143–153.
Population Action International, 1999. One in Three People Lives in Forest-scarce
Countries. Population Action International, Washington, DC.
Reid, D.G., Taylor, A.H., Hu, J.C., Qin, Z.S., 1991. Environmental-influences on bam-
boo bashania–fangiana growth and implications for giant panda conservation.
J. Appl. Ecol. 28 (3), 855–868.
from Sustainable Forestry: Guiding Principles and Practice. Kluwer Academic
Publishers, Dordrecht, The Netherlands.
Rosenstock, S.S., 1998. Influence of gambel oak on breeding birds in ponderosa pine
forests of Northern Arizona. Condor 100 (3), 485–492.
Sagar, R., Singh, J.S., 2004. Local plant species depletion in a tropical dry deciduous
forest of northern India. Environ. Conserv. 31 (1), 55–62.
Schaller, G.B., Hu, J., Pan, W., Zhu, J., 1985. The Giant Pandas of Wolong. University of
Chicago Press, Chicago, IL.
Sharma, U.R., 1990. An overview of park–people interactions in Royal Chitwan
National Park, Nepal. Landscape Urban Plan. 19 (2), 133–144.
Panda Reserve, China. Biol. Conserv. 63 (3), 231–234.
Vina, A., Bearer, S., Chen, X.D., He, G.M., Linderman, M., An, L., Zhang, H.M., Ouyang,
Z.Y., Liu, J.G., 2007. Temporal changes in giant panda habitat connectivity
across boundaries of Wolong Nature Reserve, China. Ecol. Appl. 17 (4), 1019–
Vina, A., Bearer, S., Zhang, H.M., Ouyang, Z.Y., Liu, J.G., 2008. Evaluating MODIS
data for mapping wildlife habitat distribution. Remote Sens. Environ. 112 (5),
Wolong Nature Reserve, 1998.
Nature Reserve) (in Chinese).
Wolong Nature Reserve, 2000.
in 2000) (in Chinese).
Wolong Nature Reserve, 2000.
regulations in Wolong) (in Chinese).
Nature Reserve), Sichuan Science Publisher, Chengdu, China, p. 154 (in Chinese).
Ye, Y.Q., Chen, G.J., Hong, F., 2003. Impacts of the “grain for green” project on rural
communities in the Upper Min River Basin, Sichuan, China. Mt. Res. Dev. 23 (4),
(Advances in conservation and studies on reproductivity of giant pandas in
Wolong). Sichuan J. Zool. 16, 31–33.
Zhang, P., Shao, G., Zhao, G., Master, D.C.L., Parker, G.R., Dunning Jr., J.B., Li, Q., 2000.
China’s forest policy for the 21st century. Science 288, 2135–2136.
Zhu, C., Feng, G., 2002. A Case Study on China’s Policy of Converting Steep Cultivated
Land to Forest or Grassland. China’s Forestry Press, Beijing, China.
(Master plan for Wolong
(Wolong population census
(Compilation of codes and