vh‘: v02.

.5 3"
‘21:?
‘33}.

.. . S
\3. .
it...

1.: ..

 

.aﬁunnmmnmf

C”: r436.) a

. ‘

 

 

 

.10.-
, K k!

 

 

.1... 1 .V . . A , . , , run Kt“;
4.3“.” .n 2: .3»... . 3 want .5 . I
. .. c ~ Jr. .| 1:.

.. .14.}. .7... , . . ..

 

2. (309‘

    
 

LIBRARY
Michigan State
University

 

 

This is to certify that the
dissertation entitled

BALANCING HUMAN ENERGY NEEDS AND
CONSERVATION OF PANDA HABITAT

presented by

GUANGMING HE

has been accepted towards fulﬁllment
of the requirements for the

 

 

Doctoral degree in Figlegies and Wildlife
ﬂ,iz:;:;'/v’/”w //
xii/1"", ‘. "I"
L,“ L/

 

Major Pr/ofessor’ 5 Sign tum/V

/2/ /2

Date

 

MSU is an Afﬁrmative Action/Equal Opportunity Employer

------.---.--—-.-<---—._.-o-o-v--u---—v-c--.-I-n--.—o-o---.-r-o---v---.-.-a-— -»—.-.-o-.-.—.—u-.-l-l-.-

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KrlProlecc8Pres/CIRC/Date0ue Indd

 

BALANCING HUMAN ENERGY NEEDS AND
CONSERVATION OF PANDA HABITAT

By

Guangming He

A DISSERTATION

Submitted to
Michigan State University
In partial fulﬁllment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY

Fisheries and Wildlife

2008

ABSTRACT

BALANCING HUMAN ENERGY NEEDS
AND CONSERVATION OF PANDA HABITAT

By
Guangming He

Resource extractive activities to meet the basic needs of humans have been the
leading causes of habitat degradation and thus biodiversity loss. Complex human-nature
interactions make it difﬁcult to understand the impacts of human activities on wildlife
habitat and to develop effective solutions for balancing human needs and wildlife
conservation. Integrated with modern computer technology and systems modeling and
simulation, decision support systems (DSSs) can ease this challenge. Such systems can
provide a platform for a variety of stakeholders such as conservation scientists, policy
analysts, and decision makers, to quantitatively and visually construct, analyze, and
evaluate different solution options.

To demonstrate the functionality of such systems, a WebGIS-based DSS named
WOW (Wolong Online WebGIS) has been developed. It was applied to Wolong Nature
Reserve for giant pandas (A iluropoda melanoleuca) in southwestern China. The WOW
enables various geographically dispersed stakeholders to simulate long-term spatial and
temporal dynamics of human population, human energy needs (fuelwood), human
activities (fuelwood collection and consumption), and thus the impacts of these activities
on panda habitat. This system incorporated the results from an analysis of the spatial and
temporal trends of fuelwood collection in the past three decades in the 20th century. It

also considers the effects of two conservation policies (Natural Forest Conservation

Program [NF CP] and Grain to Green Program [GTGP] implemented since 2000) on
fuelwood consumption of rural households in the reserve.

As fuelwood is a key issue of natural resource conﬂict between people and pandas,
one model was incorporated into the system to demonstrate the feasibility of thinning
plantation forests for fuelwood. It showed that forest plantations in the reserve could
sustainably provide rural households with enough fuelwood, if deliberate forest thinning
plans and conservation policy regulation were applied. Furthermore, a multi-criteria
decision making component was developed for stakeholders to synthesize and compare
scenarios or solution options. Since the system was made public, 36 of 76 registered users
evaluated four predeﬁned scenarios. The preferred scenario is to thin plantation forests
for fuelwood. Thus, we encourage reserve managers to consider this idea and incorporate
it with the on-going conservation programs such as the NFCP and GTGP as soon as
possible.

This dissertation made a novel effort to integrate traditional DSS technology and
systems modeling with spatial analyses functionality, a key component of the issue under
investigation in a web-based interactive system, which helps stakeholders search for
options balancing household energy needs and panda conservation. Besides the practical
implications of the results directly for balancing human (energy) needs and panda
conservation in the reserve studied, this dissertation has broad methodological
applications in developing Web-based decision analysis tools, assisting decision making
processes, and integrating multiple sources of data and multiple-disciplines for many of
other 2,531 natural reserves in China and many other coupled human-nature systems

worldwide.

To people who have sacriﬁced
and are willing to sacriﬁce their own well-being

for the survival of Giant Pandas

iv

ACKNOWLDGEMENTS

I would like to give my sincerest thanks to all the members of my graduate
committee, Drs. Betty Cheng, J ianguo (J ack) Liu, Angela Merti g, and Jiaguo Qi for their
long time support through the dissertation project. Particularly to my major advisor, Dr.
Liu, for his time, patience, and encouragement during this long-long eight years. I have
learned so many things personally or academically from him, which I believe will beneﬁt
me for my whole life.

I also thank all the past and current members from the Systems Modeling and
Simulation Lab and later the Center for Systems Integration and Sustainability, especially
Li An, Scotty Bearer, XiaoDong Chen, Kim Hall, Wei Liu, Ed Laurent, Yu Li, Marc
Linderman, William McConnell, Anita Morzillo, Nick Reo, HaiJin Shi, Mao-Ning Tuan
Mu, and Andres Vina. My special thanks go to the strong IT support from Jim Brown and
Blake House. Without the logistic support from Wolong Nature Reserve, I could not have
done anything. ShiQiang Zhou, Jian Yang, J inYan Huang, and a lot of nice and kind local
people who spent time answering my endless questions. A special thank you goes to
ZhiYun Ouyang of Chinese Academy of Sciences for his support in China and those
thoughtful talks with him during his visit to the lab in the summer of 2004. I would also
like to acknowledge IDV Solutions Inc., where I am currently working with, for allowing

me four weeks off in order to ﬁnish the ﬁnal writing job.

Finally, I am deeply indebted to my wife Qing Zhou and my sons Xiaochuan (Gary)
and Kevin, and my families from both my side and my wife side for their long support
and tolerance, and my four brothers for taking good care of my parents back in China.

I wish that my work deserves a “WOW” when one looks at it.

This research was made possible by grants from National Science Foundation and

National Institutes of Health to Dr. Liu.

vi

TABLE OF CONTENTS

LIST OF FIGURES ............................................................................................................ xi
CHAPTER I
INTRODUCTION ............................................................................................................... I
CHAPTER 2

SPATIAL AND TEMPORAL PATTERNS OF FUELWOOD COLLECTION IN
WOLONG NATURE RESERVE: IMPLICATIONS FOR PANDA

CONSERVATION ............................................................................................................ 10
ABSTRACT ........................................................................................................... 10
INTRODUCTION ................................................................................................. 1 1
DATA AND METHODS ...................................................................................... l3

STUDY AREA .......................................................................................... l3
DATA COLLECTION .............................................................................. 15
HOUSEHOLD SURVEY .............................................................. 16
ROAD NETWORK ....................................................................... 18
DATA ANALYSIS ................................................................................................ I8
FUELWOOD COLLECTION SITES ....................................................... I8
DISTANCE BETWEEN FUEWLOOD COLLECTION SITES AND
HOUSEHOLD LOCATIONS ................................................................... I9
RELATIONSHIPS BETWEEN FUEWLOOD COLLECTION AND
PANDA HABITAT ................................................................................... 20
RESULTS .............................................................................................................. 21
FUELWOOD COLLECTION SITES ....................................................... 21
DISTANCE BETWEEN FUEWLOOD COLLECTION SITES AND
HOUSEHOLD LOCATIONS ................................................................... 22
RELATIONSHIPS BETWEEN FUEWLOOD COLLECTION AND
PANDA HABITAT ................................................................................... 23
KNOWLEDGE ABOUT AND ATTITUDES TOWARDS FUELWOOD-
RELATED POLICIES .............................................................................. 24
DISCUSSIONS AND IMPLICATIONS ............................................................... 25
ROAD CONSTRUCTION ........................................................................ 26
ENERGY-RELATED POLICES ............................................................... 26

PANDA HABITAT CONSERVATION ................................................... 27

vii

CHAPTER 3
EFFECTS OF PAYMENT FOR ECOSYSTEM SERVICE PROGRAMS ON
FUELWOOD CONSUMPTION OF RURAL HOUSEHOLDS IN WOLONG NATURE

RESERVE, CHINA .......................................................................................................... 39
ABSTRACT ........................................................................................................... 39
INTRODUCTION ................................................................................................. 41
METHODS ............................................................................................................ 45
RESULTS .............................................................................................................. 48
DISCUSSIONS AND CONCLUSIONS ............................................................... 50

CHAPTER 4

WOW: A WEBGIS-BASED DECISION SUPPORT SYSTEM FOR PANDA

CONSERVATION ............................................................................................................ 56
ABSTRACT .......................................................................................................... 56
INTRODUCTION ................................................................................................. 58
METHODS ............................................................................................................ 60

STUDY AREA .......................................................................................... 60
SYSTEM DESIGN AND STRUCTURE .................................................. 63
SYSTEM DEVELOPMENT AND IMPLEMENTATION ....................... 64
MODELS ....................................................................................... 64

DATABASE .................................................................................. 70

INTERFACE .................................................................................. 70

RESULTS .............................................................................................................. 72
MODEL TESTING. . . .. ............................................................................. 72
SCENARIO SIMULATION ...................................................................... 72
SCENARIO EVALUATION ..................................................................... 74
DISCUSSION ........................................................................................................ 74
CONCLUSIONS ................................................................................................... 76

CHAPTER 5

THINNING FOREST PLANTATIONS FOR FUELWOOD: SEARCH FOR

SUSTAINABILITY ........................................................................................................... 96
ABSTRACT ........................................................................................................... 96
INTRODUCTION ................................................................................................. 98
METHODOLGY ................................................................................................. 101

STUDY AREA ........................................................................................ 101
FUELWOOD CONSUMPTION ............................................................. 103
FUELWOOD PRODUCTION FROM THINNING FOREST
PLANTATIONS ..................................................................................... 104
IMPLEMENTATION .............................................................................. 107
RESULTS ............................................................................................................ 108
' SENSITIVITY ANALYSIS .................................................................... 108
SCENARIO SIMULATIONS ................................................................. 108
DISCUSSIONS ................................................................................................... I 10

viii

CONCLUSIONS AND CONSERVATION IMPLICATIONS .......................... 112

CHAPTER 6
SYNTHESIS AND CONCLUSIONS ............................................................................. 137
BIBLIOGRAPHY. ..................................................................................................... 141

ix

LIST OF TABLES

Table 2.1: Eight activities not allowed in fuelwood collection ......................................... 30

Table 2.2: Three types of distances between fuelwood sites of the 19903 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) ............................................................................................................. 31

Table 3.1: Demographic, economic, and geographic characteristics of the surveyed
households ........................................................................................................... 54

Table 3.2: Effects of household characteristics on the switch of fuelwood for the three

consumption purposes to electricity .................................................................... 55
Table 4.1: Four simulation scenarios in the WOW ............................................................ 78
Table 4.2: Scenario evaluation results from 36 evaluators ................................................ 79

Table 5.1: Areas of plantations since the establishment of Wolong Nature Reserve ...... 11.4

Table 5.2: Comparisons of characteristics of Japanese larch plantations in the model and

Miyaluo ............................................................................................................. 115
Table 5.3: Initial values of parameters of stands planted in different times .................... 116
Table 5.4: Three groups of thinning scenarios ................................................................ 117
Table 5.5: Sensitivity test results for 30 years of simulations ......................................... 118

LIST OF FIGURES

Figure 2.1: The location and elevations of Wolong Nature Reserve in China. The
locations of Wolong and Gengda townships are also indicated on the map ....... 32

Figure 2.2: Temporal dynamics of fuelwood collection sites in elevation (a), slope (b) and
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 ................................................................................................................. 33

Figure 2.3: F uelwood collection site density maps of the 19703 (a), 19803 (b), and 19903
(c), and emerging fuelwood hotspots in the 19803 (b), and 19903 (c). The
density unit is points per square kilometer ......................................................... 34

Figure 2.4: Temporal dynamics of distances (ED, TD, and ND) of fuelwood collection
sites from household locations with one 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 .................................................................................... 37

Figure 2.5: Percentages of fuelwood sites of three decades (19703, 19803, and 19903)
falling in four types of habitat (Highly Suitable, Moderately Suitable,

Marginally Suitable, and Unsuitable) in 1974 ................................................... 38
Figure 4.1: Study site — Wolong Nature Reserve .............................................................. 80
Figure 4.2: Interactions among pandas, people, and policies ............................................ 81
Figure 4.3: Three-tier architecture of the WOW ............................................................... 82
Figure 4.4: A typical work ﬂow for a new user in the WOW ........................................... 83
Figure 4.5: Overview of the input, output, and analysis ﬂow of the WOW ...................... 84

xi

Figure 4.6: Algorithms of predicting the location of fuelwood collection ........................ 85

Figure 4.7: Algorithms of determining whether or not a household collects fuelwood
from natural forests .............................................................................................. 86

Figure 4.8: Screenshots of the WOW illustrating (a) the layout of the main page with four
parts, (b) models, (c) simulation results, ((1) scenario management, and (e) GIS
layers in map, and (f) menus used for navigation into other pages ..................... 87

Figure 4.9: F uelwood (a) demand and supply and (b) electricity expense dynamics of four
scenarios .............................................................................................................. 93

Figure 4.10: Impacts of four scenarios on panda habitat of four types: (a) highly suitable,
(b) moderately suitable, (c) marginally suitable, and (d) unsuitable ................... 94

Figure 5.1: Comparisons of fuelwood demand and production of the worse scenarios (a)
81-83, (b) 84-86, and (C) 87-89. Black ﬁlled squares represent fuelwood
demand and non-ﬁlled shapes are for fuelwood production. Fuelwood demands
are the same in the three ﬁgures (a, b, and c) .................................................... 1 19

Figure 5.2: Comparisons of fuelwood demand and production of the status quo scenarios
(a) C1-C3, (b) C4-C6, and (c) C7-C9. Black ﬁlled squares represent fuelwood
demand and non-ﬁlled shapes are for fuelwood production. Fuelwood demands
are the same in the three ﬁgures (a, b, and c) ................................................... 122

Figure 5.3: Comparisons of fuelwood demand and production of the better scenarios (a)
Dl-D3, (b) D4-D6, and (c) D7-D9. Black ﬁlled squares represent fuelwood
demand and non-ﬁlled shapes are for fuelwood production. Fuelwood demands
are the same in the three ﬁgures (a, b, and c) .................................................... 125

Figure 5.4: Different impacts of thinning plantation forests on highly and moderately
suitable panda habitat of the more fuelwood scenarios (3) Bl-B3, (b) B4-B6, and
(C) 87-89 ............................................................................................................ 128

xii

Figure 5.5: Different impacts of thinning plantation forests on highly and moderately
suitable panda habitat of the status quo scenarios (a) Cl-C3, (b) C4-C6, and (c)
C7-C9 ................................................................................................................... 131

Figure 5.6: Different impacts of thinning plantation forests on highly and moderately
suitable panda habitat of the less fuelwood scenarios (a) Dl-D3, (b) D4-D6, and
(c) D7-D9 ............................................................................................................. 133

xiii

CHAPTER 1

INTRODUCTION

Energy is critical to meet the basic needs of human beings. F uelwood collected in
natural forests and other ecosystems is one of the major energy sources for
numerous households worldwide, especially those in developing countries.
Approximately three billion people, half of the world’s population, are still using
fuelwood for food preparation, space heating, and other basic human needs in their
daily lives (Population Action International, 1999). However, as human populations
and household numbers are rapidly increasing (Liu et al., 2003), fuelwood collection,
together with other human activities, has led to reduction of various ecosystem
services (Leach and Meams, I988; Vitousek et al., 1997; Liu et al., 20013;
Millennium Ecosystem Assessment, 2005). For example, fuelwood collection
through cutting down trees can lead to fragmentation and degradation of wildlife
habitat (Liu et al., 2001b), reduction of wildlife populations (Aigner et al., 1998;
Hall and Farrell, 2001), and loss of biodiversity (Bhatt et al., 1994; Rosenstock,
1998; Bhatt and Sachan, 2004; Sagar and Singh, 2004). Such degradation and loss of
ecosystem services due to these activities may eventually diminish the welfare of
human beings themselves (Daily, 1997) and jeopardize the sustainable development
of human society in the long term (Millennium Ecosystem Assessment, 2005; Daily
and Matson, 2008). It is a challenge for stakeholders, such as research scientists,
policy analysts, and decision makers to balance human energy needs and wildlife
conservation and address both the needs of humans and wildlife.

Many efforts have been practiced to overcome this challenge in the past. For

example, from the demand side of fuelwood, improved cook-stoves were promoted

in India and Ethiopia (Prasad et al., 2001; Amacher et al., 2004); agroforestry
systems for mitigating the problems of fuelwood scarcity in the Montane/Savanna
ecosystems were advocated in Nigeria (Bisong et al., 2007), various renewable
energy sources and technologies such as biogas, and small hydropower and solar
energy were discussed in China, India, Nepal, and Pakistan (Rijal, 1999; Prasad et
al., 2001; Li et al., 2005). Previously scholars have focused extensively on fuel
characteristics (Vodak et al., 1986; Nygard and Elfving, 2000; Shanavas and Kumar,
2003; Nygard et al., 2004), resource availability and diversity (Samant et al., 2000),
consumption pattern (Bhatt et al., 1994; An et al., 2001; Mandondo, 2001; Tabuti et
al., 2003; Top et al., 2004), and fuelwood management (Chomitz and Grifﬁths,
2001), which are essential in understanding the conﬂicts between human energy
needs and wildlife conservation. Far fewer studies have focused on estimating
supply than demand of fuelwood. Less local biomass availability decreases the
probability of collecting fuelwood from common forests and increases the
probability of producing fuelwood on private land in Indonesia and Vietnam
(Linde-Rahr, 2003; Pattanayak et al., 2004). A recent review of the state of
economic understanding about fuelwood in developing countries concludes that
community forestry has the potential to address the gaps between household energy
needs and fuelwood scarcity (Cooke et al., 2008). Some of those studies investigated
the factors affecting fuelwood need while some focused on the variables controlling
fuelwood supply in order to tackle fuelwood scarcity and reduce the negative
impacts on wildlife habitat without signiﬁcant sacriﬁce of human welfares. However,
none of them had integratively considered both supply and demand sides of
fuelwood consumption at the household level (although household is the basic unit

of fuelwood consumption and collection) to balance human energy needs and their

ecological, social and economic impacts on wildlife conservation.

The difﬁculty of ﬁnding a balanced solution lies primarily in the complexity of
social-economic-ecological systems (Liu et al., 2007a; Liu et al., 2007b) in which
humans and wildlife compete for natural resources through human activities such as
fuelwood collection. To ﬁnd a sustainable solution of both meeting household
energy needs and minimizing the impacts on wildlife habitat, the following
questions have to be simultaneously answered. What social-economic conditions
determine fuelwood consumption of households? Where do households collect
fuelwood and what are their impacts on wildlife habitat across space and time? How
do households respond to policy interventions in terms of fuelwood collection and
consumption behaviors? Are there any fuelwood source alternatives? What are the
social, economic, and ecological impacts of alternative solutions (e.g., thinning
plantation forests as a fuelwood source alternative) on human communities and
wildlife habitat? Are there any user-friendly computer tools to assist stakeholders
answer those questions and facilitate decision making processes?

No previous efforts have been made to address the aforementioned questions in a
comprehensive fashion. A particularly well-suited site to address those questions is
Wolong Nature Reserve, China, which was designated for conserving the
endangered giant panda (A iluropoda melanoleuca) (He et al., 1996) and where
fuelwood collection by rural households has been a direct and important threat to
panda habitat in the past decades (An et al., 2001; Liu et al., 2001b). The reserve,
one of the largest panda reserves in China, is home to roughly 150 wild pandas
(approximately 10% of the wild population) (State Forestry Administration, 2005).
The pandas depend on the forests for shelter and understory bamboo for staple food

(Schaller et al., 1985). The species is solitary but could not be alone, pushed into the

high mountains and often disturbed by humans inhabiting the lower and ﬂat
bottomland. There are more than 4,500 local residents distributed in approximately
1,250 households in 2005 (Wolong Nature Reserve, 2005a). The majority of them
are farmers, living a subsistence farming lifestyle. They have traditionally collected
fuelwood from natural forests near their homes for a long time. F uelwood is used by
rural households for three purposes: cooking food for humans, cooking livestock
fodder, and heating houses in winter (An et al., 2002; Wolong Nature Reserve,
20053). From 1975 to 1998, the human population and the number of households
increased by 69% and 124%, respectively, while fuelwood consumption in the
reserve has doubled over the past two decades, consequently resulting in dramatic
habitat degradation (Liu et al., 1999a; Liu et al., 20013).

Meanwhile, the Chinese government has been implementing some conservation
policies in order to directly change social and economic behaviors of humans (e.g.,
fuelwood and electricity consumption) and indirectly protect panda habitats and
populations (Wolong Nature Reserve, 1998a; Sichuan Forestry Survey Institute and
Wolong Nature Reserve, 2000a; Wolong Nature Reserve, 20053). Some examples
since 2000 are Grain-to-Green program (GTGP, which provides subsidies to farmers
with grain and cash if they convert their cropland to forested land), Natural Forest
Conservation Program (NFCP) provides subsidies to households for monitoring one
or several forest parcels from illegal harvesting, and hydro-power plants (which
provide electricity to local residents and help them to eliminate fuelwood use)
(Wolong Nature Reserve, 20053; Vina et al., 2007; Liu et al., 20083). Due to GTGP,
more than half of cropland of households has been converted into forested land.
Almost all households participated in NFCP and are responsible for monitoring

37.5% of forests in the reserve to prevent illegal logging (the reserve staff monitor

the remaining portion of forests) (Sichuan Forestry Survey Institute and Wolong
Nature Reserve, 20003). Under these programs, fuelwood collection has been
explicitly forbidden. In the beginning of these programs (2000 — 2003) few illegal
harvesting or branches collection were seen, partially because many households still
had remaining fuelwood from previous years. As a result, it was also observed that
60~80% of the energy need for cooking human food and about 30~40% for heating
houses in winter were switched from fuelwood to electricity. Since 2004, as the
remaining fuelwood was used dead tree branches gathering has been seen and even
allowed officially permitted (Chen et 31, in review). When NFCP/GTGP programs
ofﬁcially stop, it is likely that rural households would return to their earlier practice
of cutting natural forests for fuelwood (Uchida et al., 2005). Therefore, a sustainable
fuelwood source is urgently needed in the long term to meet the energy needs of
rural households and reduce the impact of fuelwood collection on panda habitat.

A number of studies on human energy consumption and panda habitat have been
conducted in Wolong Nature Reserve. Some of them answered one or two questions
aforementioned and none of them provided a decision analysis component for
stakeholders. Using Stella 5.0 (a commercial modeling tool) and ﬁeld data from
household surveys conducted in 1998, An et al. (2001) developed 3 model to
describe the dynamics of human population and households and then predict the
fuelwood consumption of rural households for the three consumption purposes.
They found that fuelwood consumption was the functions of the socio-economic
conditions of households such as household size, with or without one or more senior
members, and cropland for corn plantations. Using SELES (a high level
programming language) Linderrnan et al. (2005) estimated the impacts of various

levels of fuelwood consumption and household creation rates on the landscape

characteristics of panda habitat for 1997-2030. An agent-based model was
developed for Wolong Nature Reserve to explore the complexity of coupled
human-nature systems (An et al., 2005). Similar to the study by Linderrnan et al.
(2005), this work also focused on the ecological impacts of demographic factors of
households on panda habitat. These models did not consider the impacts of
conservation policies on fuelwood consumption and collection behaviors. Economic
impacts of different scenarios in these models were not considered. Several
researchers have proposed to use plantation forests for human needs (Ronnback et
al., 2007; Bearer et al., 2008). For example, Bearer (2005) has studied the potential
of harvesting plantation stands varied by age for three levels of fuelwood
consumption of an average household and then predicted the amount of natural
forests that could be saved. But he did not consider the dynamics of human
population and households and did not measure the spatial and temporal impacts of
fuelwood collection on panda habitat.

Another common characteristic of these studies is lack of 3 decision making
component for stakeholders to construct, simulate and evaluate the social, economic,
and ecological impacts of simulated scenarios. Traditionally, decision support
systems (DSSs) have been widely prescribed as an analysis tool (Carlsson and
Turban, 2002; Engel et al., 2003; Phua and Minowa, 2005; Rauscher et al., 2005).
However, those systems are little accessible to stakeholders, primarily because of
complicated user interfaces targeting mainly professionals/experts (Tang et al., 2004;
Mysiak et 31., 2005; Paassen et al., 2007), high costs to deploy usually due to
expensive specialized software (Keams et 31., 2003; Adinarayana et al., 2006;
Paassen et al., 2007), lower degree of reuse and share of data and knowledge (Poch

et al., 2004; Wong et al., 2007), and without persistence storage of simulation results

and decision options allowing re-access of users. The recent advancement in intemet
technology makes a web-based DSS possible, which may address most of the
disadvantages of the traditional desktop-based DSSs.

Based on several previous studies on household population dynamics (Liu et al.,
1999c; An et al., 2001), fuelwood consumption by households (2001), and habitat
impact assessment , this dissertation aimed to analyze fuelwood collection practices
of households, study the effects of the implementation of conservation policies on
fuelwood consumption, and build 3 web-based decision support system to assist
decision makers simulating the economic, ecological, and social impacts of policy
scenarios of balancing household energy consumption and panda conservation in
their natural habitat. There are four chapters in my dissertation besides this
Introduction chapter and a brief chapter of Synthesis and Conclusions. Chapter two

analyzes the temporal and spatial patterns of fuelwood collection in the last three

decades (19703, 19803, and 19903) of the 20th century in Wolong Nature Reserve

using household survey data collected in 2002. These ﬁndings were then
incorporated into Chapter four to predict the locations of fuelwood collection. In
Chapter three, the impacts of NFCP and GTGP programs on the three purposes of
fuelwood consumption of rural households were investigated. The results were used
in Chapter four to adjust the fuelwood consumption models developed by An et al.
(2001)

Chapter four describes the system design of 3 web-based DSS called Wolong
Online WebGIS (WOW), which targets stakeholders such as research scientists,
policy analysts and decision makers to simulate the dynamics of rural human
population inside the reserve, household fuelwood demand, the spatial distribution

of fuelwood collection, the economic impacts of reducing fuelwood collection on

households, and the ecological impacts of fuelwood collection on panda habitat. In
WOW, I also created 3 multi-criteria decision making tool, which allows
stakeholders to integratively assess ecological, economic, and social impacts and to
compare and choose the balanced scenarios which enable trade-offs between panda
habitat conservation and human energy needs.

Chapter ﬁve illustrates an application of the system to search for 3 sustainable
natural resource use schema that balances household fuelwood need and panda
conservation in the reserve. Because the primary resource conﬂict between people
and pandas takes place in the forests of the reserve, I used the WOW to explore
possible strategies for achieving this balance. As such, the system helped me to
develop 3 proposal to thin plantation forests for fuelwood in the afforested areas
under GTGP and other existing reforested areas that previously experienced timber
harvesting. Simulations with a length of 30 years from 1997 to 2026 using WOW
were conducted to evaluate this feasibility. Three consumption options (more
fuelwood, status quo, and less fuelwood) each deﬁned by the switch rates from the
three fuelwood consumption purposes to electricity were considered. Fuelwood
production under different thinning scenarios (e.g., start time, cycle, and intensity)
against these three consumption options was simulated and the extents of reducing
the impacts of fuelwood collection on panda habitat were evaluated.

The multidisciplinary study incorporates ecology, wildlife biology, human
demography, systems simulation and modeling, and web-based DSS for wildlife
conservation decision making which balances household energy needs and panda
conservation. The resulting WOW could serve as a communicational tool for
stakeholders because it enables users to view scenarios proposed by others and as an

educational tool since it helps understand the complexity of coupled human-nature

systems. Thus, the decision making process would be more open and democratic,
which may increase the efﬁciency of decision making and help facilitate the
implementation of future policies. The system could facilitate a more comprehensive
understanding of complex issues, accelerate knowledge dissemination, increase
conservation awareness of the general public, and search for sustainable natural
resource uses. The methodology developed in this dissertation has broad
applications in searching for solutions balancing human needs and wildlife habitat

conservation in tightly coupled human-nature systems.

CHAPTER 2

SPATIAL AND TEMPORAL PATTERNS OF FUELWOOD COLLECTION
IN WOLONG NATURE RESERVE:
IMPLICATIONS FOR PANDA CONSERVATION
ABSTRACT
Approximately three billion people, half of the world’s population, are still using
fuelwood in their daily lives. F uelwood collection has been recognized as an
important factor in habitat fragmentation and degradation and biodiversity 1033,
especially in developing countries. Understanding spatial and temporal patterns of
fuelwood collection is fundamental to understanding human-enviromnent
interactions and designing effective conservation policies. Using Wolong Nature
Reserve for giant pandas (Ailuropoda melanoleuca) in China as an example, my
research group surveyed 200 rural households for the locations of their fuelwood
collection sites in the past three decades (19703, 19803, and 19903) 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 19703 to the 19903. 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.

KEYWORDS: fuelwood collection, spatiotemporal distribution, conservation

policy, panda habitat, Wolong Nature Reserve, China

10

INTRODUCTION

Approximately three billion people, half of the world’s population, are still using
fuelwood in their daily lives (Population Action International, 1999). The collection
of fuelwood has become important in ecological degradation worldwide, especially
in developing countries, where in many rural areas fuelwood is the sole or primary
energy source for cooking and heating (Chomitz and Grifﬁths, 2001; An et al.,
2002). Fuelwood collection through cutting down trees can lead to fragmentation
and degradation of wildlife habitat (Liu et al., 2001b), 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 31.,
2003). In order to conserve its diverse natural resources for sustainable development,
China had established 2,531 nature reserves by the end of 2007, covering more than
15% of its territory (China.com.cn, 2007). However, many of these reserves are
located in remote areas with poor economies and high human population pressures,
and various types of resource extraction are inevitable. Wolong Nature Reserve, one
of the largest reserves for protecting giant pandas (A iluropoda melanoleuca), is a

good example. Fuelwood collection is common inside the reserve and ﬂuctuated

around 7000—9400 m3/year until recently (Liu et al., 1999c). Bearer et al. (2008)

showed that panda activities in forests is reduced for several decades after timber

harvesting and fuelwood collection in this reserve. An et al. (2005) found that

. 2 . .
fuelwood collection could lead up to a 1.23 km /ye3r loss of habitat, depending on

different scenarios of socioeconomic factors. Considering the distribution of bamboo

and its periodic ﬂowering. Linderrnan et al. (2005) demonstrated that over the next

11

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 arrangement 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 perspective.

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 bamboo in the case of the giant panda (Reid et al., 1991; Taylor and Qin,
1993). It can aid in evaluating habitat suitability 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
effectiveness of previously implemented policies related to fuelwood collection
(Wolong Nature Reserve, 1998b, 2000b) could also help improve several on-going
conservation programs in China (Loucks et al., 2001; Zhu and Feng, 2003; Liu et al.,
2008b), including the Natural Forest Conservation Program (NFCP) and the
Grain-to-Green Program at Wolong. However, little has been done on gaining such
knowledge.

Therefore, this chapter aims to: (1) characterize spatial and temporal
distribution patterns of fuelwood collection in Wolong Nature 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; and (4) assess
the general relevance of the case study for fuelwood collection management in

protected areas.

12

DATA AND METHODS

Study Area
Wolong Nature Reserve was designated in 1975 to help conserve the giant panda

(MacKinnon and DeWulf, 1994). It is located in Wenchuan County, Sichuan

Province, southwestern China (Latitude 30 45'-3l 25'N, Longitude

103 52-103 24'E) (Figure 2.1 at the end of this chapter). The Wolong

Administration Bureau is responsible for both panda conservation and the
well-being of local residents (Lil et al., 2003). It reports directly to the Department
of Forestry in Sichuan Province and China’s State Forestry Administration in
Beijing (Wolong Nature Reserve, 2005b).

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 1,250 m to 6,525 m above sea level (Schaller et al., 1985). Forests
covered 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, meadows. permanent snow,
exposed rocks, roads, buildings, and water (Wolong Nature Reserve, 1998b, 2005b).
Pandas depend on the forests for cover and shelter and use the understory bamboo

(mainly Bashaniafangiana and Fargesia robusta) as staple food (Schaller et al.,

1985). They prefer slopes less than 30 , elevations between 1,500 and 3,250 m, 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) (see Liu et al., 2001b, for details on

13

habitat classiﬁcation). Approximately 4.4% of the reserve was marginally suitable,
25.6% was moderately suitable, and 5.8% was highly suitable, and the rest was
unsuitable for pandas in 2001 (Vina et al., 2007).

Timber harvesting, poaching, and agriculture were the main threats to panda
habitat and conservation. Recently, ﬁielwood collection by local residents has
emerged as an important threat. From 1975 to 1998, the human population and the
number of households in the reserve increased by 69% and 124%, respectively,
while fuelwood consumption in the reserve doubled (Liu et al., 1999b; Liu et al.,
1999c). 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 4,400 local rural residents in 970 households in two townships (Wolong
and Gengda, see Figure 2.1 for their locations). Most human settlements were in the
bottomland or on the relatively ﬂat slopes of several valleys in the reserve. In rural
China, there are three administrative units: groups, villages, and townships. Wolong
Township has nine groups in three villages, and Gengda Township has 17 groups in

three villages (Wolong Nature Reserve, 20003). The annual amount of fuelwood

consumed by each household ranged from 8 to 30 m3, depending on household size,

age structure, cropland area, and other socioeconomic conditions (An et al., 2001).
Moreover, burgeoning tourism development now may be contributing to greater
fuelwood collection 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., in press).

The household is the basic unit of fuelwood consumption (Liu et al., 1999c),
but fuelwood collection is often accomplished in winter by groups of 10—20 adult

males each from several households to increase efﬁciency and minimize risks in the

14

topographically challenging high mountains (Liu et al., 1999c). 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 to lower areas close to roads.
Finally, logs are carried and loaded into 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 removing forest canopy
(An et al., 2001). Oak (C yclobalanopsis Oerst.), maple (A cer L), birch (Betula L.),
spruce (Picea Dietr.), hemlock (T suga Carr.), larch (Larix Mill. ), and pine (Pinus L.)
are among the preferred tree species for ﬁJCIWOOd. F uelwood 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 3 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., 2001b).
Data Collection
We used historical information on fuelwood collection over the past three decades
(19703, 19803, and 19903) as well as other ecological, 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, 20003), fuelwood collection sites of the

last three decades, attitudes toward fuelwood policies, and household economy of

15

2001. Another type of data is at the reserve level, which consists of digital elevation
model (DEM), interpolated from digitalized 100 meter contours with a vertical
accuracy of less than 50 meter 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., 1999c), and road network and construction records.

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 19803
did not collect fuelwood in the 19703, 8 did not collect in the 19903, and 1 had not
collected since the 19803. Because of this, the degrees of freedom for tests of
different time combinations may be different (e. g., 385 for the comparison of the
19703 versus the 19803, and 388 for the comparison of the 19803 versus the 19903).
We recorded the most frequently visited fuelwood site for a household in each of the
past three decades on a Wolong topographic map with a scale of 1:50 000. The site
locations for all the surveyed households were then digitized and geometrically
corrected to form a geographic information system (GIS) layer in ArcGIS.

We also simultaneously investigated the attitudes of the household 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 fUCIWOOd

collection. The second regulation forbad local residents to do eight speciﬁc activities

16

during collection (see Table 2.1 at the end of this chapter). The regulations were
released in 1984 (Wolong Nature Reserve, 2000b), about halfway through our study
period. We also asked household heads whether they (I) knew about the regulations,
(2) if yes, how long they had known about them, and (3) what effects the two
regulations 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 surveyed the
household economic status of 2001 to see if relatively richer households were
different from the poorer ones in fuelwood collection activities of the 19903
(economic data of households in the 19703 and 19803 are 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, education, 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 fuelwood 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 (number 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 households

we had interviewed, each having an area of 30X30 m2 to match the habitat map.

17

With the knowledge of local guides (often participants in fuelwood collection), we
located those sites around a heavily collected spot. Old stumps (from the 19703)
were identiﬁed by newly developed seedlings from them and the seedlings also
helped us recognize tree species with the additional assistance of local botanists. We
found that on average there were 56.5 stumps left with a mean diameter of 16.2 cm
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 conifers and 56 were
broadleaves. These species were also observed in fuelwood piles during our
household interviews. Therefore, we are conﬁdent that the sites were used for
fuelwood collection.

Road network

Records of transportation system development in the reserve were kept in the
Department of Transportation in Wolong (Wolong Nature Reserve, 2005b). Road
quality, construction] improvement time, and construction goals were well
documented and road networks 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 ArcGlS 8.2 database for further analyses.

Data Analysis

F uelwood collection sites

To characterize the topography of fuelwood collection sites, average elevation,
aspect, and slope in 30X30 m2 area (0.09 ha) around the sites were derived from the
DEM. We chose this size 33 a conservative estimation of annual collection area for a
household consuming a minimum amount of fuelwood (8 m3) (An et al., 2001) and

collecting a relatively large percent (around 60%, ﬁeld observations, 2002) of trees

18

in a mixed evergreen and coniferous forest stand (most frequently visited, with an

average stock of approximately 120 m3/ha, ) (Bearer, 2005). This selection also

matches the resolution (30X30 m2) 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 kmz. Then we

identiﬁed newly emerging heavily used areas, usually centered by points with

densities of 10+ points per km2 33 fuelwood collection hotspots by comparing the

maps in different decades.

Distances between fuelwood collection sites and household locations

Because the household is a basic unit of many human activities in the reserve (An et
al., 2003), 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 distance between a
household and its fuelwood collection site. TD stands for the length of the shortest
road traveled by a household to the corresponding fuelwood site in road networks.
ND indicates the 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 household level. ND reﬂects collective interactions
between households and fuelwood collection locations at the reserve landscape

leveL

19

We used information on household expenditures of 2001 as a proxy of
economic status of the 19903. We categorized households into three groups by
annual expenditures of 2001: low (£5000 RMB, 8.25 RMB = US$ 1 in 2001),
medium (> 5 000 and $10 000 RMB), and high (> 10 000 RMB). Annual per capita
expenditures of 2001 were also classiﬁed into three categories: low ($1400 RMB),
medium (>1400 and $800 RMB), and high (> 2800 RMB). The number of
laborers in a household in 2000, derived from the 2000 population census data, was
used to group the households to test whether it affected households in terms of
fuelwood collection in the 19903 (data on the laborers in the 19703 and 19803 are not
available). We treated household members from 18 to 60 years old as laborers, and
distinguished households with two or fewer laborers from those with three or more
laborers.

Relationships between fuelwood collection and panda habitat

To examine the spatial and temporal trends of the impacts of fuelwood collection on
panda habitat, we took two measures. First, we calculated the distance (HD) from a
fuelwood site to the nearest pixel of each type of habitat on the HS] map of 1974.
Second, we derived the percentages of households collecting fuelwood in a certain
type of habitat in a given decade by overlaying fuelwood sites on the HSI map. They
may indicate temporal trends of fuelwood collection distribution in relation to panda
habitat at the reserve landscape level rather than ED or TD at the household level.

We used t tests and analysis of variance (ANOVA) to test for differences 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 different decades because aspects are circular

20

measurements (Batschelet, 1981). In addition, we calculated and reported
descriptive statistics on attitudes of local residents toward ﬁJCIWOOd policies.

RESULTS

Fuelwood Collection Sites
The temporal changes of three topographic characteristics (elevation, slope, and
aspect) of fuelwood sites are illustrated in Figure 2.2 (a, b, and c). Our results

suggest that local residents had to climb to physically challenging areas with an

average elevation > 2000 m and mean slope > 30 . 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 (19703 versusl9803 [t = 6.57, p < 0.0001]
and 19803 versus 19903 [t = 3.47, p = 0.0003]) show that site elevation signiﬁcantly
increased over time (on average, 100—150 m per decade) (Figure 2.23). The slopes
of the three decades were not signiﬁcantly different (F = 2.73, p = 0.066) while they
increased only from the 19703 to the 19803 at the signiﬁcance level of 0.05 (t = 2.33,
p = 0.016) (Figure 2.2b). The Mardia—Watson—Wheeler test shows that there were
no signiﬁcant differences among the aspects of fuelwood sites during the three
decades studied (W = 5.92, p = 0.21) (Figure 2.2c).

Density maps of fuelwood sites used during the 19703, 19803, and 19903 are
plotted in Figure 2.3 and clearly indicate clustering of fuelwood collection across the
reserve landscape. Several fuelwood hotspots had emerged over time in the reserve
(Figure 2.3). We identiﬁed Area A as a hotspot in the 19803 by comparing the maps
in the 19703 and 19803 (Figure 2.3b). Similarly, two hotspots, Area B and Area C,
attracted many households for fuelwood in the 19903 (Figure 2.3c). Furthermore,
linking road construction information with these hotspots shows that road

development might have contributed to their emergence. Roads near Area A were

21

constructed from a trail to improve local access to markets with support from the
United Nations’ World Food Programme for panda conservation in the early 19803
(W olong Nature Reserve, 2005), but extended so far into the forests that they clearly
facilitated fuelwood collection in Area A. Roads near Area B were originally
developed for mining in 1987 and assisted many households in fuelwood collection
in the 19903 (Wolong Nature Reserve, 2005). The quality of the main road across
the reserve, from the east side to the lower southwest comer and constructed for
commercial timber logging in the 19603 before the establishment of the reserve, was
greatly improved in the early 19903 (it was paved, which was rare in western China
during that time). These roads fragmented the reserve, and facilitated household
fuelwood collection considerably. People living far away from the road could easily
travel to Area C, which had high forest stock and was rarely used as 3 fuelwood site
before the mid—19903.

Distances between F uelwood Collection and Household Locations

Besides climbing higher into the mountains, as mentioned above, local residents also
traveled greater distances (TD) to fuelwood sites, because sites became increasingly
farther away from households (ED) over time (Figure 2.4). On average, local
households had to reach out an extra 20—50 m farther (ED) or travel 50—80 m more
(TD) yearly to ﬁnd good fuelwood sites. Although the distance difference seems
small, it often occurred in steeper and higher 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 (Figure 2.4).

22

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 19903 (Table 2). Presumably, this is because fuelwood collection
was accomplished by groups of households rather than by individual households.
Another reason might be that unavailability of fuelwood was the same for all
households, whether they were poor or rich, with more or fewer laborers.
Relationships between Fuelwood Collection and Panda Habitat
The percentage of fuelwood sites in highly suitable panda habitat of 1974 increased
from 6% to 21% during the three decades studied (Figure 2.5). The majority 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 (Figure 2.5). Surprisingly,
no fuelwood collection appeared in any areas of marginally suitable habitat. The
reason might be because marginally suitable habitat is characterized by steep slopes
and high elevations (Liu et al., 1999c). From the 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 m, 132 m, and 105 m for the 19703,
19803, and 19903, respectively) and farther from unsuitable habitat (90 m, 127 m,
and 145 m for the 19703, 19803, and 19903, respectively) to ﬁnd good sites for
fuelwood over time. A

To further demonstrate the spatial trends of fuelwood collection impacts on
panda habitat, we calculated the distance between a household and its nearest habitat

pixel of a given type. The locations of households were relatively stable. Our results

23

show that for highly suitable habitat, the average distance between households and
the 1974 habitat (153.43 2t 131.59 m)was signiﬁcantly shorter than the one between
households and the 1997 habitat (330.86 :t 264.58) (t = 12.80, p < 0.001). Similarly,
for moderately suitable habitat, the mean distance corresponding to the 1974 habitat
(113.14 :t 99.11) was signiﬁcantly shorter than the one corresponding to the 1997
habitat (236.35 2t 169.36) (t = 13.86, p < 0.001). Because the data for panda habitat
in the 19803 were not available, we only analyzed the habitat change of the
fuelwood sites from the 19703 to the 19903 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 19703 had become unsuitable and
the rest had no change. One-quarter of sites that were moderately suitable in the
19703 had become unsuitable and the rest were unchanged. In total, the quality of
18.8% of fuelwood sites in the 19703 had been reduced by the 19903. To summarize,
the continual collection of fuelwood in the 19703, 19803, and 19903 had degraded
habitat quality and affected the spatial distribution of panda habitat; primarily,
suitable habitat had been pushed farther away from households.

Knowledge about and Attitudes toward Fuelwood-related Policies
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 regulation. 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 policies 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 among those who followed the regulations, only approximately half

24

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 of regulations on fuelwood collection time, 43% of
household heads 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 regulations) and collected more small trees (93% of
households who followed regulations). Thus, regulatory policies and the
unavailability 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.

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 the majority of tree fellings

occurred in areas of highly suitable and moderately suitable panda habitat rather
than areas of marginally suitable and unsuitable habitat. Fuelwood collection
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 pandas. The implications of our results for policy are discussed

below.

25

Road Construction

Our study suggests that road systems facilitated fuelwood collection 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-agricultural activities in the past should have been closed to prevent
illegal use for resource extraction.

Energy-related Policies

Over time, local residents have experienced increasing difﬁculties in collecting
fuelwood, involving tedious trips to steep and remote mountains. Energy is a crucial
issue for both human welfare and panda conservation. This is becoming even more
important 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 forest 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 the majority of households to reduce
fuelwood consumption since 2001 by switching to electricity (Wolong Nature
Reserve, 2005b). The average electricity consumption per household increased from

800 kWh in 2000 to 2800 kWh in 2003 while fuelwood consumption per capita

declined from 1.4 m3 to 0.3 m3 during the same period (Wolong Nature Reserve,

2005b). A new, so-called ecohydropower plant began operation in October of 2002,

26

but the price of electricity, a widely proposed substitute (An et al., 2002; Wolong
Nature Reserve, 2005b), 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 0.05 RMB could save

approximately 15 km2 of panda habitat over a period of 30 years (An et al., 2006).

With substantial income losses when NFCP ends, the households 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 (Ye et al., 2003; Zhu
and Feng, 2003). However, 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 recommend that the majority of steep cropland
returned to vegetation should be designated speciﬁcally for growing fuelwood
forests, providing both energy resources and ecological ﬁrnctions in the reserve. By
continuing to use fuelwood. local residents (of three minority ethnic groups: Tibetan,
Qiang, and Hui) may preserve their traditional lifestyles and cultures (N gugi, 1988;
Mahiri and Howorth, 2001).

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

27

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 difﬁcult and relatively large areas, but the Wolong
Administration Bureau had limited staff for resource monitoring and protection
(Wolong Nature Reserve, 2005b). Second, the tragedy of the commons (Hardin,
1968) is applicable in Wolong: collecting fuelwood from forests is free, except for
the time it takes (Liu et al., 1999c), and, 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 ﬁrelwood needs.

Third, some of the local residents’ basic needs (e.g., energy) were not well
addressed (Sharma, 1990): many potential alternatives (e.g., biogas and wind/sun
power) were unavailable, were expensive (e. g., coal, because of transportation), or
had other negative environmental effects (e. g., greenhouse gases from burning
charcoal and coal). Although the eight hydropower plants in the reserve had a total
capacity of 33 960 kW in 2003 (Wolong Nature Reserve, 2005b) and local residents
were willing to switch to electricity (An et al., 2002), the people could not afford to
buy it and most of the electricity was sold to cities. It was found that at the average
price of 0.13 RMB/kWh in 1999, a price decrease of 0.05 RMB would have doubled
the number of households switching to electricity (An et al., 2002).

We suggest that resources or efforts should be shifted to monitor fuelwood
collection in the areas of highly suitable and moderately suitable habitat. Currently,
the farther an NFCP parcel is from the household assigned to monitor it, the more

subsidies the household receives. To improve the efﬁcacy of NFCP’s investment,

28

household 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 have negative impacts on ecosystem health and services
(Bengtsson 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 very few were
illegally harvested. This will also give enough time for the reforested trees in the
Grain-To-Green Program to grow for fuelwood.

Understanding spatial and temporal patterns of fuelwood collection will not
only help researchers scale up their evaluations of the impacts of fuelwood
collection on panda habitats from the household level to the reserve landscape level,
but also provide fundamental knowledge for reserve managers to make informed and
effective decisions, and evaluate and adjust policies. For example, different
conservation priorities could be set for areas with different types of habitat quality
affected by fuelwood collection. Furthermore, 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.

29

Table 2.1: Eight activities not allowed in fuelwood collection

 

Number

Description

 

l
2

OOQG'JIA

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
Reserve, 1998b).

30

80:08 0:: 30:0: 88800

00230 8 8>& 080 80:00.: :08 80.: 0:028:00 9:08 0:: 0:0 .058 05 m0 308 58.: 0:: 8 830% 0m? 08 O 080 .m .< 80:00.: 8:08
0:9 .8008 008:: m0 <>OZ< 83-080 :0 028-8 0:: m: 028: 80:88 05 :80 U :0 .m .< 80:00.: 0:: .3 8:08 0:: m0 <>OZ< .33
-080 (:0 08:80.8 0:: m: 88:80ch 0:“ 028:0 8:88 0:: .308 08008 05 8 8030300 800830 0:: m: 028: 8:88 0:: 080 anew
0:: 8 80:80.6 0:: .«0 :88 0:: m: 88:80qu 0:: 028:0 80:88 0:: .308 E: 0:: 8 80808588 008380 :0 09¢ :08 con: 08 Z
.8888 .m.D 3800 m: .0888 H mg : a

 

A8088

24.836 8082: 5.832 .
0.20880: “8888 80::

 

0:0 85 0003.08 800250:

88:8 @883 28:82 233% 2.3.303. 3383 835 889% 8838 5.088052

 

 

 

$0.886 Amodvmmd Ammdvmvd A8808 £20880:

:80 8% 000203 800250:

$2.08; 5303: $8383 63:88 535% 88888 88:82 084:8.” 883323355
Ed: 3 5.836 5.880

A8088 £20880:

 

N :80 8:8 0003—08.: 80080:

5202.: 82.004: 5822” E85: 38038 585: 82:23 20308
808820 8802035 Om

 

 

mN NW oowmA oowmlooi oovW coo 3A coo Elogm oooww
8% 00:0 8
A8208: :0 8:825 08868098 888 .808 20:00:05 mam—am H a .D
O m 8 0.868088 20:00.80: :80;
<

 

ADV ooom 8 20:08.0: :08 8 808:8 :0 80:88 080 Am: 53. m0 0.8208088 888 :08 20:08.0: A<v Sow .«0 08868090

20:00.50: :80: 05 x: 008008 020:080: :0 80:80: :80 mag: 05 :0 8% 000.200.: 80.50: 808880 m0 899 008,—. ”Nd 050,—.

31

Figure 2.1 The location and elevations of Wolong Nature Reserve in China. The

locations of Wolong and Gengda townships are also indicated on the map

 

wolong Township]

   
 

 

Canda Township

Elevation (111)

l I <2_()()()

“ 2.000 - 3.000

- 3.000 — 4.000
- 4.000 - 5.000
- >5.000

  

 

 

 

32

Figure 2.2 Temporal dynamics of fuelwood collection sites in elevation (a), slope (b)
and 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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a
‘g~ 2L800 ; ~ , ,9 , e A w —7W 4* fees
5 l W
- 20400 i T
E w l
'u * T
g, :Looo
<2 1,600 4 ~~ ~~~ . ~—v~T - i
1970s 1980s 1990s
Time
b
- 60 T 7 74* 4*
En 50 1
V 40 1 T T T
j? 30 ‘ stir
M
20
E .01
< 0 l r
1970s 1980s 1990s
Time
C
E; 360 l8 9* r
g, 270 4
g sol
3 T
5. 90 i T
g 0 l 7 Ci: ,7 1 ,
<
1970s 1980s 1990s
Time

33

Figure 2.3 Fuelwood collection site density maps of the 19705 (a), 19805 (b), and 19903
(c), and emerging fuelwood hotspots in the 19803 (b), and 19905 (c). The density unit is

points per square kilometer

 

 

 

 

 

 

 

 

 

 

  

Legend

[:3 Boundary /

Main road /

~~—-- Secondn 13' roads /

 

Site density

l 0 - 5

-5-10
0 1.5 3 6Km -l0+

 

 

 

 

 

 

 

34

Figure 2.3 continued

b

 

 

 

 

 

 

 

 

 
  
 

 

V

9.." ﬁt”
'.
54’

Legend

[:3 Boundary

—— Main road

       

 

—— Secondary roads
Site density

[j 0 - 5

5 - 10

o 1.5 3 6Km -“,+

 

 

 

 

 

 

 

35

Figure 2.3 continued

 

 

 

 

 

 

 

   
 
 

\’\.

      

I)?
vie/{rt

My

Legend

:3 Boundary

— Main road

'
-¢\

 

—— Secondary roads
:7." Site densitV

ﬂ '

— l V ,, l 0 - 5

[:1] 5-10
0 1.5 3 6Km -10+

 

 

 

 

 

 

 

36

Figure 2.4 Temporal dynamics of distances (ED, TD, and ND) of fuelwood collection
sites from household locations with one 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

 

 
   

 

 

 

 

 

 

 

 

 

 

 

 

4,000 i 9 ++
A 3,000l E11970s
E ‘ I1980s
V
q, l1990s
3 2000 _,
a , .
H
.2
G

1,000

0 l 1
TD ND
D'ntance Type

37

Figure 2.5 Percentages of fuelwood sites of three decades (19705, 19805, and 19905)
falling in four types of habitat (Highly Suitable, Moderately Suitable, Marginally

Suitable, and Unsuitable) in 1974

 

 

 

 

 

 

 

 

 

 

80
' 019705

60‘ ‘ B19805
d.)
an . I19905
g l 1
= 40 ‘
O
U 1‘»
2:
m 20 I Hi"

0 L 1 ”4 T 1

 

 

 

 

Highly Suitable Moderately Suitable Marginally Suitable Unsuitable

Habitat Type

38

CHAPTER 3

EFFECTS OF PAYMENT FOR ECOSYSTEM SERVICE PROGRAMS ON
F UELWOOD CONSUMPTION OF RURAL HOUSEHOLDS IN WOLONG
NATURE RESERVE, CHINA

ABSTRACT
Payment for ecosystem service (PES) programs are becoming popular worldwide in
attempts to resolve conﬂicts between natural resource conservation and human
development. Whether human society accepts them socially and economically and
thus adjusts their behaviors largely decides the ecological success of those programs.
Although natural resources are the major source of conﬂicts between conservation
and development and household is the basic unit of decision making, relatively little
effort has focused on understanding the effects of such policy interventions on
resource-extractive behaviors at the household level. To ﬁll this knowledge gap, I use
Wolong Nature Reserve for Giant Pandas, China, as an example. The reserve is home
to approximately 150 wild pandas and 1,000 households. Fuelwood collection has
been one of the main driving factors of panda habitat degradation and loss in the past,
and two PES programs—the Natural Forest Conservation Program (N PCP) and the
Grain-to-Green Program (GTGP)—have been implemented since 2000. In 2002, my
research group surveyed 200 rural households regarding their energy consumption
switch from fuelwood to electricity after the implementation of the two PES policies.
We also investigated what social, economic, and geographic factors of households
affect the switch. We found that two-thirds of the fuelwood required by households
for cooking human food was replaced by electricity, and the switch was not

signiﬁcantly related to any household characteristics; one-third of the household

39

fuelwood demand for heating houses was met by electricity, and household income
and age of household members were the signiﬁcant driving factors; and households
were least apt to substitute electricity for the fuelwood needed for cooking livestock
fodder, and the percentage of cropland converted in GTGP and household size
signiﬁcantly affected the decision to make this switch. NFCP and GTGP together
reduced the energy dependence of rural households on fuelwood from 100% to
approximately 64%. Even with substantial economic support from PBS programs, it
is hardly possible for rural households to switch most of their energy needs to
electricity, especially for heating and cooking livestock fodder. Thus, other
ecologically ﬁiendly and economically sound energy alternatives should be
considered. One promising fuelwood source would be plantation forests, and local
governments should encourage tree plantations for both household use (e. g.,

fuelwood) and ecosystem service (e. g., erosion control, biodiversity conservation).

KEYWORDS: fuelwood collection, payments for ecosystem services, Wolong Nature

Reserve, China, wildlife conservation, Grain to Green Program (GTGP), Natural Forest

Conservation Program (NFCP), fuel switch

40

INTRODUCTION

The damages of global environmental crises (such as global warming, ﬂooding,
deforestation, and desertiﬁcation) to economies and human society have been
increasingly recognized by many governments worldwide, even in deve10ping countries.
It is encouraging that in order to prevent such damages in the future, some of those
nations have initiated and implemented large-scale conservation programs (Sayer et al.,
2004; Weyerhaeuser et al., 2005; Chokkalingam et al., 2006; Engel and Palmer, 2008),
for example, payments for ecosystem services (PBS) in China (Daily and Matson, 2008;
Liu et al., 2008b).

China’s large-scale national conservation programs after the disastrous ﬂooding
along Yangtze River in 1998 are excellent examples (Trac et al., 2007; Liu et al.,
2008b). Pilots of the Natural Forest Conservation Program (N FCP) began immediately
following the ﬂooding. Its short-term goals (1998—2000) were to eliminate or reduce
timber harvesting from natural forests and create alternative employment for traditional
forest enterprises. The medium-term goals (2001—2010) are to construct and protect
forests for ecological beneﬁts and to increase the capacity for timber harvesting ﬁ'om
plantation forests. The long-term goal (2011—2050) is to restore natural forests and meet
domestic demand for timber in plantation forests (Liu et al., 2008b). Started one year
after NF CP, the Grain-to-Green Program (GTGP) in China aims to increase vegetative
cover by 32 million ha by 2010 through the conversion of cropland on steep slopes to
forest and pasture, with an additional "soﬁ" goal of afforesting a roughly equal area of
wasteland (Liu et al., 2008b). Pending successful completion, it could represent a

10—20% increase in China's national forest area and a 10% decrease in current

41

cultivated area (Bennett, 2008). Although ecological impacts are rarely or too early to
be conﬁrmed by ﬁeld observations (Trac et al., 2007; Wunder et al., 2008), social and
economic successes have often been documented at large scales in both the scientiﬁc
literature (Uchida et al., 2005; Shen et al., 2006; Xu et al., 2006a; Bennett, 2008;
Goldman et al., 2008; Liu et al., 2008b) and government reports (such as from the State
Forestry Administration). However, more studies on the effects of such programs at
ﬁner scales, such as the household level, are needed to understand the effects of these
programs at larger scales (county, provincial, and national levels), adjust the policies,
and improve implementation (Weyerhaeuser et al., 2005; Xu et al., 2006b; Trac et al.,
2007; Turner et al., 2007).

Wolong Nature Reserve of China is an excellent site for studies to reduce this
knowledge gap. The reserve, with approximately 36% forest cover in 2001 (Vina et al.,
2007), has nearly 4,000 plant species and more than 2,200 insect and 124 animal
species, including the endangered giant pandas (A iluropoda melanoleuca) (Wolong
Nature Reserve, 1998b). From a national survey conducted in 2000, 143 wild pandas
were believed to cohabit the reserve with 4,498 people (63% of which are Tibet) in 968
households in 2003 (State Forestry Administration, 2005; Wolong Nature Reserve,
2005b). The reserve provides important ecosystem services such as ﬂood control and
biodiversity to the region. Both NF CP and GTGP have been initiated in the reserve
since 2000 (Wolong Nature Reserve, 2005b). The collection of fuelwood, the major
energy source of rural households, is one of the main threats to maintaining forest cover
and wildlife habitat and population. If these two programs are to be successful in

Wolong Nature Reserve, the local rural households must switch from fuelwood to

42

electricity with the economic support from the programs, assuming other energy sources
such as plantation forests are not considered (Bearer, 2005). Naturally, the next
questions to ask are: How much do these two programs trigger households to stop
fuelwood collection and switch to electricity? and What factors affect this decision
process at the household level?

One of the most popular theories about household energy choice is the “energy
ladder” model (Leach, 1992; Heltberg, 2005; Arnold et al., 2006). It predicts that with
increasing afﬂuence, households transition from relying solely on biomass to a mixed
use of kerosene, coal and charcoal, and ﬁnally the cleanest fuels, liquid petroleum gas
and electricity. This model prescribes income to be the only impact factor. Recently,
several other factors have been considered (Heltberg, 2005; Farsi et al., 2007). These
factors can be categorized into three groups: demographic, economic, and geographic.

Household demographic factors include education level and sex of household head
(Farsi et al., 2007), labor force, household size, number of educated people in the
household (Heltberg, 2004; Chen et al., 2006), and the highest education level of
household members (Heltberg, 2004; Madubansi and Shackleton, 2007). The economic
factors used most often in the literature are household income (Heltberg, 2004, 2005;
Farsi et al., 2007; Madubansi and Shackleton, 2007), per capita expenditure (Heltberg,
2004), fuel price (Heltberg, 2004; Farsi et al., 2007; Hiemstra-van der Horst and
Hovorka, 2008), and cropland size and wealth (Chen et al., 2006). For geographic
factors, researchers have considered access to fuel sources (Masera et al., 2000),

distance to forest (Chen et al., 2006), and regions (Heltberg, 2005).

43

The majority of fuel-choice or transition studies only consider fuel for cooking or
total fuel usage, partly because most studies are conducted in tropical regions, where
fuel needs for heating houses in winter may be negligible. However, fuelwood in many
other regions is often consumed for three purposes: cooking human food, heating
houses, and cooking livestock fodder. The consumption for each purpose may vary by
geographic region. Different factors may play different roles in decisions to switch fuels.
Most fuel-switch research has focused on the status quo of fuel consumption (Masera et
al., 2000; Heltberg, 2005; Farsi et al., 2007) while little effort has linked the switching
process to conservation interventions. Another common characteristic of such studies is
that they concentrate at large scales such as multiple countries, provinces, and regions.
These studies may not be so helpful to a relatively small area, such as Wolong Nature
Reserve, because many macrofactors are similar or the same for all households; for
example, fuel price, cultural preference, scarcity of fuelwood, law enforcement, and
availability of electricity. These factors were not considered in this study. One
exception is that Madubansi and Shackleton (2007) examined the changes in fuelwood
use and selection in ﬁve villages of South Africa and found that even with a policy of
6kWh per month of free electricity, over 90% of households still used fuelwood for
cooking and heating. Their study did not associate socioeconomic factors with the
fuelwood consumption changes of different fuel-use purposes.

The goals of this chapter are to (1) estimate the effects of two PES conservation
policies (NFCP and GTGP) on switching from fuelwood to electricity for three

fuelwood consumption purposes at the household level; (2) identify social, economic,

44

and geographic factors affecting the process; and (3) provide recommendations for
adjusting NFCP and GTGP for panda conservation.

It is my hope that by providing a better understanding of how these PES programs
affect rural household decisions, the programs can be adjusted not only to be more
effective but to ensure the sustainability of both human society and panda habitat in the
reserve. These insights may be also applicable to other areas covered by the PES
programs and similar regions worldwide.

METHODS

Located in Sichuan Province, southwestern China (Latitude 30 45'—3l 25'N, Longitude
103 52'—103 24'E, Figure 2.1) (He et al., 1996), Wolong Nature Reserve was

established in 1963 and expanded to its current size of 2,000 km2 in 1975. In 1980,

Wolong Nature Reserve became a member of the Man and Biosphere Programme’s
World Network of Biosphere Reserves and, in 2006, was inscribed on the World
Heritage List (UNESCO World Heritage Centre, 2006).

The reserve is administrated by Wolong Administrative Bureau and consists of two
townships: Wolong and Gengda (Lii et al., 2003). Four major ethnic groups—Han,
Tibetan, Chang, and Hui—constitute the rural population, most of which are farmers.
Households use fuelwood as their main energy source for cooking human food and
livestock fodder and heating in winter. Fuelwood needs for these three purposes consist
of 18.3%, 45.7%, and 36.0% of the total energy demand, respectively (An et al., 2001).
Many potential alternatives are unavailable (e. g., biogas and wind/sun power),

expensive (e. g., coal, because of transportation), or have negative environmental effects

45

(e. g., charcoal and coal). Electricity is the ideal substitute for fuelwood because the
reserve’s eight hydropower plants have an adequate capacity—33,960 kW in 2003
(Wolong Nature Reserve, 2005b). However, before 2000 many households consumed
little electricity, primarily for lighting and electric appliances (e. g., TV and radio)

because of low quality and the relatively high price (An et al., 2002). The annual

amount of fuelwood consumed by each household ranged from 8 to 30 m3, depending

on household size, age structure, cropland area, and other socioeconomic conditions

(An et al., 2001). This fuelwood consumption has caused a signiﬁcant reduction in the
quality and quantity of panda habitat (Liu et al., 1999c; An et al., 2005; Linderrnan et al.,
2005).

Almost 100% of households in the reserve participated in both NF CP and GTGP
(Wolong Nature Reserve, 2005b). Farmers who participated in GTGP received 2,250 kg
of grain and 300 yuan (1 USD = 6.8 yuan in 2008) per ha of converted cropland per
year. More than 50% of cropland in the reserve (367.3 ha) has been enrolled in GTGP.

A group of households was assigned a forest parcel for collectively monitoring illegal
harvesting and each household was subsidized with an average of 1,210.04 i 14.39

yuan per year, which was 19.24 :t 0.96% of the yearly household income in 2001.

Direct cash gains from GTGP contributed a very small portion (1.51 :1: 1.39%, Table 3.1)
to the total household income, but the cash value of the grain equaled 17.14 :h 1.16% of
the average household income (the agricultural beneﬁts of planting corn, potato,
cabbage, and other crops if the land was not converted were not deducted). However,
households may save rather than consume the income gains (Ravallion and Chen, 2005).

Even if they decide to consume, they might not spend money to reduce resource

46

extractive activities (e. g., consuming less fuelwood and more electricity) but in other
areas such as education and health. Administering NFCP and GTGP differently from
other conservation programs, China has dispatched about 60,000 specially trained forest
police to emphasize program enforcement (Wang et al., 2007). Together with the
monitoring efforts by households, the existence of such a forest policy force in the
reserve (approximately 50 people) has helped monitor the implementation of those
policies in remote areas and has led to almost no occurrence of illegal forest harvesting
and poaching (Wolong Nature Reserve, 2005b). Partially because of this, households
may have to switch from fuelwood to electricity to meet their energy needs.

We interviewed 200 households (about one-ﬁfth of the total rural households in the
reserve) in the summer of 2002. Those households were sampled based on a random
sampling method stratiﬁed by three community organizational levels (township, village,
and group) using the population census data of the year 2000 (Wolong Nature Reserve,
2000a). We collected household demographic and socioeconomic data, including
cropland returned to forest through GTGP and economic gain from GTGP and NF CP
(Table 3.2). Particularly, we surveyed the percentages of energy consumption switched
from fuelwood to electricity for each of the three purposes (cooking human food,
heating, cooking livestock fodder) after the implementations of the two conservation
programs in the reserve. Before 2000, little electricity was consumed for the three
purposes but for some household electric appliances such as radios and TVs because of
the low quality of electricity. We chose the household heads or their spouses as

interviewees because they are usually the decision makers of household affairs. Of the

47

200 households interviewed, seven were excluded from analysis because their
households did not complete all the questions.

We summarized and reported descriptive statistics of demographic, economic, and
geographic factors affecting households’ decisions to switch from fuelwood to
electricity. We also used t tests and analysis of variance (ANOVA) to test for
differences of the percentages of fuelwood switched among the three consumption
purposes. Furthermore, we applied linear regression models to estimate the effects of
these household characteristics (independent variables) on the percentages (dependent
variables) of fuelwood for each of the three consumption purposes switched to
electricity.

RESULTS

The demographic, economic, and geographic characteristics of the surveyed households
are summarized in Table 3.1. Surveyed households had an average size of 4.44
members in 2002 but decreased to 3.86 in 2006. Approximately one-third of the
households had at least one senior member older than 60 years. Our respondents had
low levels of education with a mean of 2.96 years. Most households had constructed or
remodeled their houses more than a decade before 2002. The average per capita
cropland was 2.12 mu (1 ha = 15 mu) but more than half of the households (54.72%)
were enrolled in GTGP. The average per capita household income was 2,416.21 yuan in
2001. NF CP directly contributed 19.24% of the household income while the
contribution of GTGP was negligible. The average household elevation of 1,862.53 m
conﬁrms that most are located in higher mountains, which may require more energy for

heating in winter. The mean distance from a household to the main road was 528.04 m.

48

Approximately 46% of surveyed households were from Wolong Township, located in
the southwestern part of the reserve (Figure 2.1).

The effects of the two PES programs on rural households’ switch from fuelwood to
electricity varied for the three fuelwood consumption purposes. Households switched a
major amount (66.92 i 2.42%) of fuelwood need for cooking human food to electricity.
The switched percentage of fuelwood need for cooking livestock fodder is the least

(22.62 :1: 2.61%). A little more than one-third of fuelwood need for heating in winter

(37.05 i 3.12%) was met by consumption of electricity. The one-way ANOVA (F2576

= 68.41 , p < 0.0001) demonstrates that households switched signiﬁcantly different
percentages of fuelwood consumption for the three purposes. Further, one-tail paired t
tests indicate that the switch percentage of fuelwood need for cooking human food is
statistically higher than it is for heating (t = 9.69, p < 0.0001), which is higher than the
switch percentage for cooking livestock fodder (t = 4.39, p < 0.0001).

Effects of the demographic, economic, and geographic characteristics of
households on the switch of ﬁielwood for each of the three consumption purposes under
the two PES programs GTGP and NFCP are presented in Table 3.2. The three groups of
household attributes explained 26% and 23% of the variance of the percentages of
switching from fuelwood for heating (F = 5.27, p < 0.001) and for cooking livestock
fodder (F = 4.53, p < 0.001), respectively. We found that no geographic attributes of
households were signiﬁcantly related to the fuelwood-to-electricity switch for these
three fuelwood consumption purposes. Although none of the household characteristics

signiﬁcantly affected households to switch fuelwood need for cooking human food to

electricity (R2 = 0.06, F = 0.96, p = 0.484), economic factors affected by the two PBS

49

programs signiﬁcantly impacted the switch for the other two fuelwood uses. For the
switch-to-electricity percentage of fuelwood for heating, per capita income (t = 3.215, p
= 0.002) was positively and signiﬁcantly associated and age of house (t = -2.484, p =
0.014) was negatively and signiﬁcantly associated. The percentage of cropland returned
to forest through GTGP positively inﬂuenced the switch of fuelwood to electricity for
cooking livestock fodder (t = 3.186, p = 0.002). Household size affected that switch
negatively (t = -3.5097, p = 0.077).
DISCUSSIONS AND CONCLUSIONS
Our results indicate that rural households in the reserve had switched different portions
of fuelwood need for different consumption purposes to electricity under the two PES
programs NF C P and GTGP. A higher percentage of the fuelwood need for cooking
human food was switched, followed by heating and cooking livestock fodder. Overall,
approximately 36% (18.3% X 66.92% + 36.0% X 37.05% + 45.7% X 22.62% = 35.9%)
of fuelwood dependence was reduced. Multiple fuel sources would coexist for long
terms, and this ﬁnding is consistent with the literature (Masera et al., 2000; Arnold et al.,
2006). F urthermore, the regression analysis showed that socioeconomic factors directly
related to the two PES programs together with some demographic characteristics of
households signiﬁcantly impacted the process of switching from fuelwood to electricity
for two major consumption purposes while geographic factors were not signiﬁcant.

The difﬁculty of switching fuelwood for cooking livestock fodder to electricity
may be explained by two reasons. One is that the amount of fuelwood need for this
purpose was big (45.7% of the total fuelwood demand). The other is that proper

cookware for using electricity did not exist to efﬁciently cook livestock fodder. These

50

two reasons did not exist for cooking human food, which may explain why households
chose to buy electricity to replace most of the fuelwood for that need.

Households need a relatively large amount of fuelwood for heating (36.0% of the
total fuelwood demand). House quality and structure may play more important roles
when switching fuelwood to electricity. We observed that many old houses in the
reserve were built with stones and soil and new ones with bricks and cement. More
afﬂuent households are more apt to upgrade their houses. Therefore, it is likely more
difﬁcult for households living in older houses (often poorer) to switch more fuelwood to
electricity for heating. Fuelwood for heating may be the only case where people can
sacriﬁce their own welfare to reduce the need and cost if they do not have alternatives,
and this is true in the reserve. Thus, more investment from the two PES programs can
increase the switch to electricity in two ways. Investments can be made to help poor
households remodel their houses and increase the per capita payments, which will
improve heating efﬁciency and encourage households to consume more electricity for
heating.

An et a1. (2001) showed that the amount of cropland for com (20—30% of the total
acreage) was positively correlated to the fuelwood consumption for cooking livestock
fodder. This may explain why more GTGP-converted cropland correlated to a higher
switch from fuelwood to electricity for cooking livestock fodder. Although this is
positive for conservation, the potential impacts on economy (e.g., pigs may grow slower)
and culture (the local residents prefer pork from pigs fed by cooked food) are not clear

and deserve more research.

51

One limitation of the study may be that it was conducted only one year after NF CP
and GTGP were initiated in the reserve. This is reﬂected in that household
characteristics explained merely a little more than 20% of the variance, although the
relationships between some households attributes affected by the programs and the
switch of fuelwood to electricity for heating and cooking livestock fodder were
signiﬁcant. Our survey may be better interpreted as short-term responses of rural
households to these programs. Another uncertainty is that during the beginning of the
implementation many households still had a substantial amount of fuelwood left from
previous years. Although our study demonstrates that rural households responded
signiﬁcantly differently for the three fuelwood consumption purposes, more recent data
are necessary to enhance the understanding of the effects of the PES programs on fuel
switch of rural households.

This analysis can provide some useful implication for PES policy design and
implementation and for panda conservation. To encourage households to switch more
fuelwood to electricity for heating, the reserve needs to consider more measures; for
example, distributing more economic beneﬁts from ecotourism to the local communities
(He et al., in press) and off-farm economic activities (Zhang et al., 2008) to improve the
income of households so they can buy more electricity or remodel their houses for
higher energy efﬁciency. Realizing the challenge of completely substituting fuelwood
with electricity and in order for the two PES programs to succeed in the long run, we
may consider that an appropriate balance needs to be achieved between amelioration of
the negative consequences of fuelwood on the environment and improvement of human

well-being. This is especially important in the areas where fuelwood has been

52

traditionally consumed and treated as part of the culture of the rural communities

(N gugi, 1988; Mahiri and Howorth, 2001; Wolong Nature Reserve, 2005b). Community
forestry and plantation forests have the potential to address fuelwood scarcity and
energy crisis (Bearer, 2005; Cooke et al., 2008). One option would be the use of GTGP
plantation forests, currently few of which were designed and are suitable for fuelwood,
and the government should encourage households to plant trees for both household uses
(e. g., fuelwood and fumiture) and ecosystem services (e. g., erosion control and
biodiversity conservation). These efforts will likely sustain the achievements of NFCP
and GTGP and eventually improve panda habitat by reducing fuelwood collection from
natural forests in the reserve. The results from this study may also have broad
implications for other human-populated protected areas in China and other parts of the

world.

53

Table 3.]: Demographic. economic, and geographic characteristics of the surveyed

 

 

 

 

households
Cate or Variable name Descri tion Mean
g y p (Standard deviation)
Demographi hh_size Household size 4.44 (1.54)
c
has_senior Whether household 0.32 (0.47)
has at least one senior
member older than 60
years (1: yes, 0: no)
head_edu Education (years) of 2.96 (1.00)
household head
Economic house_age Age of house (years) 16.05 (13.50)
capita_income Per capita income 2416.21 (2099.44)
(yuan)
gtgp_income_pct Percentage of income 1.51 (1.39)
from GTGP
nfcp_income_pct Percentage of income 19.24 (13.31)
from NFCP
gtgp_cropland_pct Percentage of 54.72 (17.64)
cropland returned in
GTGP
capita_total_cropland Per capita cropland 2.24 (1.63)
(mu, 1 ha =15 Mu)
Geographic loc_code 1: Wolong; 0: Genda 0.46 (0.50)
elevation Elevation of house 1862.53 (223.65)
(meter)
dist2road Distance from house 528.04 (696.85)

to the main road
(meter)

 

54

Table 3.2: Effects of household characteristics on the switch of fuelwood for the three

consumption purposes to electricity

 

Independent
variables

F uelwood consumption purposes

 

Cooking
human food

Heating

Cooking

livestock fodder

 

Intercept

hh_size
has_senior
head_edu
house_age
capita_income
gtgp_income_pct

nfcp_income_pct

gt gp_cropland _pct

capita_total_cropla
nd

loc_code
elevation

dist2road

71 .6696(O.174)
-2.5670(0.206)
1.5795(0772)
3.0229(0.246)
-O.1246(0.511)
0.0006(O.720)
-1.6675(0.499)
-O.1136(0.670)

0.1963 (0.193)

-2.1318(0.283)
-1.4961 (0.913)
-0.0006(0.984)

-0.0034(0.57l)

R2 = 0.06
F = 0.96
p = 0.484

98.6604(0.102)
-O.l336(0.954)
-1.7085(0.784)

3.2673(0.273)
-0.5382(0.014)

0.0065(0.002)
-2.6947(0.340)
-0.0688(O.821)

0.2294(0.l83)

-3.2908(O.148)
0.8715(0.956)
-0.0413(0.234)

-0.0009(0.900)

R2 = 0.26
F=527
p < 0.001

84.4945 (0.101)
-3.5097(0.077)
1.2881 (0.809)
-3.7281 (0.143)
-0.0834(O.652)
0.0018(0.288)
-3.1658(O.189)
0.2802(0.282)

O.4667(0.002)

-2.4721 (0.202)
10.5706(0.430)
-0.0316(0.284)

0.0083(0.158)

R2 = 0.23
F=4s3
p < 0.001

 

Note: Numbers in ﬁ'ont of the parentheses are coefﬁcients and numbers inside are
p-values for the independent variables or intercepts in the same row; the last row
indicates the other results of linear regressions for the fuelwood consumption purposes
in the same column.

55

CHAPTER 4

WOW: A WEBGIS-BASED DECISION SUPPORT SYSTEM
FOR PANDA CONSERVATION

ABSTRACT
A daunting challenge for conservation scientists is to assist various geographically
dispersed stakeholders in understanding the dynamics of complex human-nature
systems and making informed decisions. Traditionally, decision support systems (DSSs)
have been widely prescribed as an analysis tool to help the decision-making process.
However, several characteristics of these DSSs (e. g., desktop-based, high costs,
complicated interfaces) have limited the access by multiple stakeholders such as
scientists, reserve managers, policy makers, non-proﬁt conservation organizations, and
the general public. With the development of Internet technology, web-based DSSs have
emerged with a universal user interface but usually are loosely integrated and often lack
spatial components. During my research, 1 developed a web-based DSS named WOW
(Wolong Online WebGIS) that is integrated with geographic information systems (GIS),
system modeling, and decision analysis for wildlife conservation. We applied the
system to Wolong Nature Reserve, one of the largest reserves in China for conserving
the endangered giant panda (A iluropoda melanoleuca). Although the main purpose of
WOW is to help with increasingly complex decisions of general wildlife conservation,
the methodology developed here has broad applications in integrating multiple sources
of data and multiple disciplines to tackle natural resource management issues in many

other coupled human-nature systems worldwide.

56

KEYWORDS: decision support system, systems modeling and simulation,

multicriteria decision analysis, wildlife conservation, web, giant panda, China

57

INTRODUCTION

Wildlife conservation challenges often arise in coupled human-nature systems(Liu et al.,
2007b). Human actions may have economic and social impacts on human society as
well as ecological effects on wildlife habitat and population across time and space (An
et al., 2005; An et al., 2006; Liu et al., 2007a). It is crucial for various geographically
dispersed stakeholders (e. g., the general public, conservation mangers, decision makers,
interested groups, and conservation scientists) to understand the dynamics of such
complex systems. The keen competition between conﬂicting interests of stakeholders
makes decisions and decision support even more challenging. This challenge is also
being exacerbated by lack of appropriate decision support tools that evaluate these
conﬂicts integratively.

Many applications have been developed to tackle this challenge rooted in such
complex systems. However, the stakeholders have limited access to them. The primary
reasons may include complicated user interfaces targeting mainly professionals/experts
(Tang et al., 2004; Mysiak et al., 2005; Paassen et al., 2007), high costs to deploy
usually due to expensive specialized software (Keams et al., 2003; Adinarayana et al.,
2006; Paassen et al., 2007), lower degree of reuse and share of data and knowledge
(Poch et al., 2004; Wong et al., 2007), and without persistent storage of simulation
results and decision options allowing re-access by users.

With recent technological innovations in computer sciences, the Internet is
remarkably popular in both developed and developing countries. For example, as of
December 31 , 2007, the Internet population in mainland China was estimated to be

over 210 million, only 5 million away from replacing USA as the world’s largest wired

58

nation (China Internet Network Information Center, 2008). It is worth noting that 40%
of China’s new lntemet users in 2007 were from rural areas (Schwankert, 2008). This
provides a good user base for Web-based applications in China. WebGIS-based
decision support systems (DSSs) are becoming the ideal solution to the aforementioned

challenge (Bhargava et al., 2007b). Though researchers have claimed that “the Web is
now the platform of choice for building DSS ” (Bhargava and Power, 2001), few

applications that address conservation are seen in the literature.

WebGIS-based DSS is advantageous to its classical standalone forerunners in
several ways. The availability of the systems on the World Wide Web (Web) naturally
expands accessibility of the systems and this may potentially generate public awareness,
a critical factor in the successes of conservation practices (Shunula, 2002; Solh et al.,
2003; Dobson, 2005). It also improves effectiveness of the system by a universal
interface allowing almost zero learning cost (Tang et al., 2004; Bhargava et al., 2007a).
No further investment (i.e., for often expensive GIS and database software) for end
users except a “free” intemet browser, e. g., Netscape or lntemet Explorer, is another
beneﬁt. Domain knowledge and expertise can also be easily distributed and shared
across time and space by users of the systems without extra costs. All these advantages
together may promote democratic decision making and encourage public participation
in wildlife conservation.

Based on previous research efforts (Liu et al., 1999c; An et al., 2001; Liu et al.,
2001b; An et al., 2002; An et al., 2005; Linderman et al., 2005; An et al., 2006), we
have developed a new Web-based DSS, Wolong Online WebGIS (WOW). Different

from most existing Web-based applications with only data query and/or map browsing

59

services (Keams et al., 2003; Lu, 2004; Sante et al., 2004; Mathiyalagan et al., 2005),
this system is additionally equipped with spatial analysis functionalities and simulation
models, which are necessary to demonstrate and illustrate the complicated
interrelationships among people, pandas, and policies in our study site—Wolong Nature
Reserve, China.

The rest of this chapter introduces the study area, system design, and
implementation of WOW. I also present the results and a discussion. Finally,
conclusions and future research directions are presented.

METHODS
Study Area

Wolong Nature Reserve is located in southwestern China (see Figure 4.1). It was

designated as a nature reserve in 1963 with an area of 200 km2 and expanded to 2,000

km2 in 1975, with about 10 % of China’s wild pandas living in the steep mountains

from 1,200 m to 6,500 m in elevation (Figure 4.1). Pandas’ survival relies on the quality
and quantity of habitat that consists of forests as shelter and bamboo as their staple food
source. They prefer ﬂat terrain or areas with gentle slopes for ease of movement. Mixed
conifer and broadleaf forests with understory bamboo located in the elevation range of
2,250—2,750 m are the most suitable habitat for giant pandas (Schaller et al., 1985).
However, the reserve is also home to more than 4,500 rural residents in about

1,100 households (Wolong Administration Bureau Department of Social and Economic
Development, 2006). Most of the local residents are farmers, living in a subsistence
economy and extracting “free” forest resources through agriculture expansion, road

construction, housing building, Chinese medicine gathering, and fuelwood collection.

60

Human activities are the major factor threatening panda habitats in the reserve (Liu et
aL,l999c)

The interactions among people, pandas, and policies are complex (Figure 4.2).
Rapid human population increase, changes in household structure and size, and fast
economic development have accelerated the panda habitat degradation (Liu et al., 1999c;
Liu et al., 2001b; Vina et al., 2007). This negative impact has endangered the panda
population to the brink of extinction, and triggered the government to implement many
conservation-oriented policies historically and continually. For example, three programs
have been initiated in the reserve since 2000: hydropower plants provide electricity to
local residents to help them eliminate fuelwood use, the Grain-to—Green Program
(GTGP) provides subsidies to farmers with grain and cash if they convert their cropland
to forested land (by 2003, 75% of cropland, or 560 of 742 ha, had been planted with
trees or bamboos), and the Natural Forest Conservation Program (N F CP) provides
subsidies to households for monitoring one or several forest parcels from illegal
harvesting (e. g., fuelwood collection) (Vina et al., 2007; Liu et al., 2008b). These
policies are aimed to directly change human social and economic behaviors and
indirectly protect panda habitats and populations. For instance, under serious
enforcement of and with direct grain or cash subsidies from these policies, many
households dramatically reduced fuelwood consumption by switching to electricity after
2000 (Wolong Nature Reserve, 2005b). It was estimated that about 60—80% and
30—40% of the energy needs for cooking human food and heating houses in winter,
respectively, were switched from fuelwood to electricity (see chapter 3). Another

notable change is that a number of local households, especially those located in lower

61

elevations, started to feed their pigs uncooked fodder during periods that varied from
one month to a whole year. This accounts for a reduction of 20—30% in fuelwood for
cooking livestock fodder (see chapter 3). These changing processes and their potential
impacts on panda habitat are characterized with nonlinearity and delay and may end up
with unexpected results (Liu et al., 2007a). It is essential for decision makers to
understand the dynamics of these interactions in the long run.

We consider fuelwood collection, a main human activity affecting panda habitat
and population, to be a key link among people, pandas, and policies in the reserve. In
WOW, models were developed to simulate the dynamics of rural human population at
the individual and household levels, calculate fuelwood consumption based on
demographic and social factors of households, estimate fuelwood production if
harvesting plantation forests, and predict the locations of fuelwood collection and the
costs of using electricity for each household. WOW can also evaluate the dynamic
impacts of fuelwood collection on panda habitats at the reserve landscape level given
different policy scenarios. Based on the multicriteria decision-making (MCDM)
methodology (Malczewski, 1999), we also developed a policy scenario evaluation
component to help decision makers assess and compare different policy scenarios and
then choose the best based on their own ecological, economic, and social preferences.

WOW is web-enabled and accessible to stakeholders from anywhere and at any
time. Stakeholders for panda conservation in Wolong may include local residents
spatially dispersed in the reserve, reserve managers and government ofﬁcials of various

administrative levels (e. g., county, province, and nation) geographically distributed in

62

China, as well as scientists, NGOs. and the general public geographically isolated
worldwide.

System Design and Structure

The WOW system consists of four components: browsers, Web server, Web application
server, and database in a three-tier framework as illustrated in Figure 4.3. Users interact
with the system through the Web as a universal interface in a browser. They send
requests from browsers to the Web server. The Web server ﬁlters those requests for
WOW and passes them to a dedicated instance of the Web application server. The
application server parses them and invokes the models, which then interact with the GIS
server and the database server to retrieve spatial and non-spatial data and execute
necessary spatial and non-spatial operations. These two latter servers ﬁnish most
computational tasks in the models. When this step is done, the application server
transfers the resulting outputs into the format of HTML and sends them through the
Web server back to browsers for users to view.

In WOW, most computationally intensive operations such as spatial analyses and
system simulation are done in the server side located remotely (e. g., the Web server and
the Web application server). In this way, domain models and expensive GIS software
are shared. It also eases data management and promotes information sharing. This
design may reduce system performance because more transportation between clients
and servers is required. Thus, several endeavors were made to use computing
capabilities in the client side (i.e., users’ browsers) as well, e. g., mouse positions and
actions on GIS maps were collected by J avaScript, and a system menu implemented in

J avaScript facilitates navigation between Web pages.

63

We chose Internet Information Services 6.0 (118) as our Web server, Tomcat 5.0 as
the Web application server, ESRI ArcGIS Server 9.0 as the GIS Server, and Oracle
Express 10g as the database server. The system supports most standard browsers, such
as lntemet Explorer and Netscape. All components are free under GNU licensing terms
or bundled with hardware or are discounted for educational uses. WOW was developed
and deployed as a part of the Web application server. The system was integrated in a
Windows server 2003 environment and is accessible at http:/A3anda.csis.msu.edu/wow.
System Development and Implementation
Java was chosen as the primary development language because it enables
interoperability and is free (Palmer, 2000). System models work tightly with the
ArcGIS server, providing access to spatial data and GIS functionalities. User
information and policy scenarios are stored in databases, allowing later evaluation and
re-access from different users. The Web interface was implemented using J avaServer
Page (J SP) technology based on J avaServer Faces (J SF) framework (Armstrong et al.,
2005). Model results are visualized in the form of graphs, tables, and maps. A typical
work ﬂow for a new user of the system is illustrated in Figure 4.4. System development
and implementation are detailed below.

Models

The core of WOW is the model base consisting of seven models. Figure 4.5 illustrates
an overview of the input, analysis processes, data ﬂow, and the output of these models.
The human population dynamics model and fuelwood consumption model (An et al.,
2001; An et al., 2002; An et al., 2006) are brieﬂy discussed ﬁrst. The other ﬁve models

are then described in more detail.

64

consumption model (An et al., 2001; An et al., 2002; An et al., 2006) are brieﬂy
discussed ﬁrst. The other ﬁve models are then described in more detail.

The human population dynamics model simulates yearly events of individuals in
each rural household in the reserve since 1996, including birth, death, marriage, and
migration (out of or into the reserve because of education or marriage) (An et al., 2001).
Household dynamics (e. g., creation, dissolution) are also considered (An et al., 2001).
After this simulation, the fuelwood consumption model predicts the amount of
fuelwood each household consumes each year based on the household size and structure,
and other socioeconomic factors such as the amount of cropland belonging to the
households (An et al., 2002; An et al., 2006). To simulate the years after 2000, the
effects of NF CP and GTGP can be considered if users prefer to do so. A user can '
specify a switch rate from fuelwood to electricity for each component of fuelwood
consumption (cooking human food, cooking livestock fodder, and heating houses).
These switch rates are applied to each household in the reserve while the rates for the
years before 2000 are calculated based on the socioeconomic conditions of speciﬁc
households (An et al., 2002).

Reforestation for fuelwood model: Though there are plenty of plantation forests,
mainly of Japanese larch (Larix kaempferi (Lamb) Carr), from afforested cr0p1ar1ds or
reforested clear-cut timberland in the reserve (Wolong Nature Reserve, 2005b), which
are believed to not be suitable habitat for pandas (Bearer, 2005; Bearer et al., 2008),
households are not allowed to harvest them for fuelwood by the present regulations.
This model estimates the production of fuelwood if afforested larch plantations are

thinned in the reserve given different combinations of thinning intensity, starting year,

65

and thinning cycle. A larch plantation growth model under thinning was derived from
the literature. Households are assumed to collect fuelwood from thinning plantation
stands rather than in natural forests if this model is included in the simulation. If
fuelwood is not needed from natural forests, no fuelwood collection site is allocated for
the households by the fuelwood site distribution model.

Economic impact of fuelwood collection model: The less fuelwood households
collect from natural forests or plantation forests, the more electricity they consume and
the more cash they need to pay. Using standard heating conversion equations for fuels
(Forest Products Laboratory, 2004), we convert a certain volume (in cubic meters) of
fuelwood substituted by electricity to the equivalent amount (in kWh) of electricity and
calculate the costs of electricity (in USD) for each household. Finally, the model reports

the total annual electricity expenses as

N
ElectricityExpensei =( X FWSubstitutediJ') x 0.7 x 574,000 / 3,340 x Electricity Pr ice / 7.0, [4-1]
1

where ElectricityExpensei is the total electricity expense (USD) of the year i,

FWSubstitutedij is the volume of fuelwood (cubic meters) switched to electricity in the

year i for the household j (j is from 1 to N, the total number of households), 0.7 is the
average density (ton/m3) of fuelwood, 574,000 is the net heating value of a unit of
fuelwood (Btu/ton), 3,340 is the net heating value of a unit of electricity (Btu/kWh),
ElectricityPrice is 0.10 RMB/kWh in or before 2000, 0.18 RMB/kWh after 2000, and
7.0 is the currency exchange rate from RMB to USD (RMB/U SD).

F uelwood site distribution model: To analyze the impacts of fuelwood collection

on panda habitat, we need to know where households collect fuelwood. To achieve this,

66

we adopted an empirically based methodology from our ﬁeld observations and
household surveys conducted in 2002 about farmers’ behaviors of fuelwood collection
(Neto, 2003). The methodology is depicted in Figure 4.6.

F uelwood collection in natural forests is often accomplished yearly in the winter
by groups of 10—20 people in the reserve. Most group members are neighbors, friends,
or relatives. They shared fuelwood sites to increase efﬁciency and minimize risks in the
rugged high mountains. Households usually do not change areas for ﬁrelwood over time
unless fuelwood stock runs low enough in those areas (the model used 20% as the
threshold based on our household surveys). Otherwise, they may ﬁnd another spot close
to the previously used one. Thus, we deﬁned these areas as potential fuelwood
collection areas (PFAs) which are topographically similar polygons, each associated
with a group of households that collect fuelwood collaboratively. The 280 NFCP
parcels were delineated (Sichuan Forestry Survey Institute and Wolong Nature Reserve,
2000b) so that they are topographically similar to facilitate forest monitoring. A more
recent survey also shows that local households collected dead trees or branches
(illegally), often in their own parcels, after NF CP was started (Chen et al. unpublished).
Thus NFCP parcels could be treated reasonably as PF As in WOW.

The initial PFAs of the 200 households surveyed are the NFCP parcels where they
collected fuelwood in the 19905. The initial PFAs of the remaining households were
determined to have been those of their nearest-neighbor households that we surveyed.
Newly formed households use the same PFAs as their parental households. In the
simulation, a randomly selected point (pixel) inside a PF A polygon is assigned to the

household associated with the PF A as its fuelwood collection site in a given year if the

67

household needs to collect fuelwood from natural forests. The algorithm to determine if
a household collects fuelwood from natural forests is described in Figure 4.7.

Habitat impact assessment model: After the fuelwood collection site (point) was
derived for each household, the type of habitat where fuelwood collection occurs was
determined by overlaying ﬁJelwood sites on a map of the panda habitat suitability index
(H81) in the habitat impact assessment model with a 30X30-meter resolution. The
resulting maps were used for further landscape-level spatial pattern analyses of human
activities and panda habitat. For example, the total areas of habitat affected by fuelwood
collection grouped by habitat type were calculated as a preliminary indicator of the
impact on panda habitat in WOW.

Multicriteria decision-making model: Generally speaking, all conservation actions
can directly and indirectly have ecological, economic, and social impacts on human
communities and/or wildlife habitat and populations (Rossi et al., 2008). Therefore,
WOW was designed to allow each decision maker to integratively evaluate a scenario
(deﬁned by the parameters from all the models) from the three criteria (Ecological,
Economic, and Social). Each decision maker has three weights assigned to one criterion.
Each is ranged from 0 to 1 and totaled to l; the higher the weight is, the more important
a scenario is from the corresponding criterion. Based on the output from the scenario
simulation and their personal preferences, decision makers then assign a score to the
scenario for each criterion. The score is in the scale 0—10; the higher the score is, the
less negative impact the scenario has on the corresponding criterion and the more the
scenario is preferred from the corresponding criterion. The most commonly used

decision rule, Simple Average Weighting (SAW), of the multicriteria decision-making

68

methodology (Malczewski, 1999) was implemented to summarize these scores, as

brieﬂy explained below.

The score Sij (0—10) that a scenario 1 receives from a user j is calculated as
n th
Sij = k£1 (Wk X Skij) , where Wk is the weight ofthe k criterion the userj

perceives, n (= 3 in our case) is the number of criteria the user j has, and Skij is the score

the user j gives to the scenario i based on his/her criterion k. The total performance

in
score 5i (0—10) of the scenario 1 is given as Si = ( ZSij) / m, where j represents the
j=l

.th . .
j user who evaluates the scenano, and m 15 the number of users who evaluate the

scenario i. Based on the Si, the scenarios can be ordered. The higher the score is, the

better the scenario is perceived collectively by evaluators or stakeholders and
integratively from social, ecological, and economic perspectives.

In this prototype, WOW provides information mainly on the ecological, economic,
and social impacts of scenarios. WOW is aimed to provide a platform and support the
stakeholders in organizing and synthesizing information for the issue under
consideration. Some other impacts are left to stakeholders to explore (Mysiak et al.,
2005; de Anguita et al., 2008): the possible human health improvement if switching
from fuelwood to electricity, the monetary value and cost of habitat restoration, the
effects of human activities on panda behavior, and the economic effects of NFCP and

GTGP on governments and rural households. WOW users or scenario evaluators may

69

express their opinions on these impacts by adjusting the scores to the scenarios when
evaluating these impacts.

Database

To simulate fuelwood needs and collection and their impacts on panda habitat, WOW
processes and manipulates two major types of data: spatial and non-spatial. They are
stored in different places with appropriate formats. Demographic data on 4,413
individuals and 893 households in 1996 (An et al., 2001; An et al., 2006) are stored in
two text ﬁles as initial states of the models and loaded into the computer memory when
the system is initialized. Replicates are made in computer memory every time a new
simulation starts. Locations of households and their associated PF A polygons and HSI
maps from 1997, derived from remotely sensed imagery (An et al., 2001), exist as ﬁles
and are managed as layers in the ArcGIS server. They are accessible to models through
Java codes. Information about WOW users (name, password, email address, etc.), their
preference weights of evaluation criteria, their proposed scenarios (consisting of model
parameters), scenario simulation results including number of households, number of
individuals, fuelwood demand, fuelwood produced from reforestations, annual
electricity expense, and areas of impacted habitat by type, and scenario evaluations
(evaluation scores, see Table 4.2) are stored in an Oracle Express database and updated
on-the-ﬂy. Some examples of intermediate data are information on the amount of
fuelwood consumption of each household, newly born babies, new households, and
fuelwood collection sites. They may be kept temporarily or cleared when a new
simulation is initialized to save storage space on the server side.

Interface

70

To facilitate efﬁcient and effective use of the system, WOW’s interface was designed
for desktop application, following the Model-View-Controller (MVC) pattern widely
practiced in software engineering (Sun Microsystems Inc., 2002). Users can accomplish
most tasks (e.g., model parameter input, systems simulation, scenario evaluation and
management) and view results in a single web page after logging onto the system
without navigating to other web pages. Each action updates only the corresponding
part(s) on the page. This page is divided into four parts and the screenshot in Figure
4.83 demonstrates this.

Users can change the parameters of models and simulations and start a simulation
in the upper left part (Figure 4.8b). The simulation results are shown as colored lines
with legends in the lower left comer (Figure 4.8c). Most key tasks (updating criteria;
creating, deleting, updating, loading, and evaluating a scenario) related to scenario
management can be done in the lower right part (Figure 4.8d). In the upper right part
(Figure 4.8e), household locations, H81, and other complementary information of the
system are shown and users can operate most basic GIS functionalities (e. g., zoom in
and out, pan, identify, layer control, scale, overview). Some complementary information,
such as user name and currently loaded scenario 1D, is shown above this part under the
page title (Figure 4.8c). Other supporting documentations such as help are deployed on
other web pages as well. All the web pages are driven by a menu located at the top of
the screen (Figure 4.81). Because of the universal interface of Web and online help (e. g.,
when the mouse pauses on a button for a while, the functionality of the button will show
up in a separate, small window), little time is needed for users to learn. The basic GIS

capabilities can help users become familiar with the study area and the research issues

71

(Figure 4.8e). Further information about the system models and the study site is
provided as web links of referenced literatures. which can be opened in a separate
browser window.

RESULTS

Model Testing

Most of the models used in WOW have been veriﬁed and validated in published
literature. For example, the human population dynamics and ﬁrelwood consumption
models were developed and validated in An et al. (2001; 2005; 2006), the reforestation
for fUCIWOOd model in Li et a1. (2000) and Ma et al. (1998), and the habitat suitability
map used in the habitat impact assessment model in Liu et al. (1999c). The equations
and parameters used for the economic impacts of the fuelwood collection model have
been tested by the US. Forest Service (Forest Products Laboratory, 2004). The
fuelwood site distribution model was validated by comparing the simulated fuelwood
collection locations with the ones from surveys conducted in 2002 (He et al., in review).
The results show that all surveyed sites in the 19905 were inside the fuelwood collection
buffer zone deﬁned as the maximum polygon that contains all the predicted locations
during the same period. Approximately 64.5% of surveyed fuelwood sites in the 19905
were within 120 m of predicted locations during the same period, 21.6% were within
120—180 m, and the remaining 13.9% were within farther distances. In summary, the
models work well and WOW’s integration with these models gives us reasonable
conﬁdence for later decision analysis.

Scenario Simulation

To demonstrate the functionalities and capabilities of the WOW system, four scenarios
were simulated for a period of 20 years (1997—2016) with 20 replicates. Parameters
used for each model in the four scenarios are shown in Table 4.1. We intended to show
the effects of NFCP and GTGP and thinning plantation forest stands for fuelwood, so
human population parameters were set the same for those scenarios. Scenario A is the
baseline assuming NF CP and GTGP were not implemented and without thinning
plantation forests for fuelwood. Scenario B simulates fuelwood consumption under the
impact of NF CP and GTGP. Scenario C simulates fuelwood consumption under the
impact of NF CP and GTGP and fuelwood supply from harvesting plantation forest
stands. Scenario D assumes higher rates (than ones used in scenarios B and C) of
switching from fuelwood to electricity, and determines whether fuelwood supply from
plantation forests meets the consumption need. Some of the results on fuelwood need
and supply are shown in Figure 4.9.

Compared to Scenario A, Scenario B showed that implementation of NFCP/GTGP
may reduce approximately 30% of the fuelwood consumption of households but
households still need to cut natural forests to meet the remaining 70% of fuelwood.
Considering reforested stands as the source of fuelwood, Scenario C may reduce 50% of
fuelwood needs from natural forests. Harvesting plantation forests could provide a
substantial portion of fuelwood needs. The existing larch plantations could supply
enough fuelwood for all households so that no natural forests are needed any more if
higher rates of switching to electricity could be realized, possibly by greater investment
of NFCP and GTGP after year 9 (Scenario D). Consequently, a large amount of panda

habitat could be preserved from reduced ﬁrelwood collection in natural forests (Figure

73

4.10). Compared to Scenarios A and B, Scenario C saved more than 50% of affected
habitat of each type (Highly, Moderately, Marginally suitable and Unsuitable) during
most of the simulated years. Since in Scenario D fuelwood needs of households are
reduced (e. g., by 50% of the needs in Scenario C, Figure 4.9a), the areas of affected
habitat of all types are gradually decreased to zero (year 9). Meanwhile, we should
notice that the economic burden on households to pay for electricity is increasing
(Figure 4%). The annual electricity expense per capita in Scenario D doubles the ones
in Scenarios B and C, which double the one in Scenario A.

Scenario Evaluation

We made the system public through several listservs of professional organizations such
as the Ecological Society of America (ESA), the United States Regional Association of
the International Association for Landscape Ecology (US-IALE), and the Society of
Conservation Biology (SCB) in early October 2008. As of November 1, 2008, 76 users
worldwide (eight countries in four continents) had registered in the system. Thirty-six
of them have evaluated these four scenarios and more than 50 new scenarios have been
constructed and simulated. The statistics of the scenario evaluation are shown in Table
4.2. Scenario D is preferred by most evaluators, followed by Scenarios B, C, and A. The
ranks of these four scenarios may change when more users evaluate the scenarios.

DISCUSSION

This preliminary WebGIS-based DSS was designed and developed to demonstrate the
capability of integrating Web, GIS, DSS, and system modeling and simulation for
coupled human-nature systems (Liu et al., 2007a; Liu et al., 2007b). Several models in

the system were simpliﬁed and could be reﬁned. For example, in the fuelwood site

74

distribution model, fuelwood sites were randomly chosen in the PF A polygons in order
to improve the computational performance. This may not represent the real-world
situation because the pixel with more fuelwood stock in the PFA is usually favored.
However, it could be enhanced by storing the fuelwood stock of the pixels in the PF A
polygons and assigning the pixel with the maximum amount of stock to the household
that collected fuelwood in the PF A.

The resulting dynamics of the numbers of households, household size, and
fuelwood consumption and supply can be illustrated for only one scenario at one time.
The current implementation does not allow comparisons of the results across scenarios
simultaneously. This weakness could be overcome by extending the current
visualization functionality. By design, a new plug-in (e. g., a Java class) can be added
and the corresponding J SP could be easily modiﬁed for the plug-in without compiling
the source codes of the core models or making major changes to the system structure.

The simulation of four scenarios showed that the proposal of harvesting larch
plantations could provide substantial fuelwood or even more than the total need after
year 10. The ideally best scenario (Scenario D with the switch probability to electricity
from fuelwood for cooking, heating, and feeding pigs is 90%, 65%, and 35%,
respectively) is reasonable based on our current ﬁeld observations in the reserve. With
NFCP and GTGP, and economic structure adjustments (especially because of
ecotourism development, W. Liu at al. unpublished data), local household income may
increase gradually and more electricity would be consumed in the future. We believe

that it is feasible because a complex program such as NFCP has been implemented and

75

enforced successfully by the local government (Wolong Nature Reserve, 2005b; Vina et
al., 2007; Liu et al., 2008b).

The multicriteria decision-making method SAW used to evaluate decision
alternatives in this study is utility-based, dominated in the multicriteria based-GIS
applications (Malczewski, 2006). The outranking relation-based method, the other of
the two main families of methods, may be useful since it allows the scenario evaluators
to consider qualitative evaluation criteria for which preference interval ratios have no
sense, limit the compensation between evaluation criteria, and in general, require
limited information from the decision makers (Chakhar and Mousseau, 2008).

The four scenarios were introduced as a demonstration of the capabilities of WOW.
Other scenarios (e.g., by changing demographic factors) can be proposed as they are
being constructed by users worldwide in the search of the best policy interventions. The
proposal of new and different scenarios is an indication of collaborative
decision-making processes using WOW as a platform.

CONCLUSIONS

We presented a WebGIS-based DSS WOW to tackle a complicated wildlife
conservation issue in coupled human-nature systems. With this platform, various
geographically dispersed stakeholders can simulate the dynamics of human population
and fuelwood needs, predict the locations of fuelwood collection, and then assess the
impacts of fuelwood collection on the human community and panda habitat in Wolong
Nature Reserve. Besides this, it demonstrates how such a spatial DSS can provide a set
of solution spaces on which decision makers can focus their discussions and make

collective choices assisted by multicriteria decision-making methodology. It can also

76

enable decision makers to have a better view of the problem and to test different
scenarios by varying the objectives, constraints, and parameters of the models.

The functionalities and capabilities were shown as well. The system presented here
can potentially help stakeholders understand the dynamics of complicated coupled
human-nature systems across different temporal and spatial scales. The demonstrated
methodology can facilitate and assist decision-making processes because it integrates
multiple sources of data and multidisciplinary knowledge within a universal
user-friendly framework. The development done here also provides a good foundation
that allows future extension and reﬁnement of operational systems. This methodology
can be used for similar issues in many other parts of the world.

Furthermore, the system enables group decision making. Although conservation
problems encompass multiple stakeholders and their preferences should be combined to
choose the best option, in reality, decisions are often made by one or few interest groups,
such as central governments. Usually community members, NGOs, and the general
public do not have a say or their inputs are rarely considered in the decision-making
process. WOW provides an open platform for them to participate in the decision making.
WOW can also be used as an educational tool for the general public to facilitate

understanding of the dynamics of complex human-nature systems.

77

Table 4.1: Four simulation scenarios in WOW

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Scenario
Model Varrable A B C D
Fertility
(children/woman) 2'5 2'5 2'5 2'5
Marriage age (years) 22 22 22 22
College attendance rate 1.92 1.92 1.92 1.92
(%)
Human Female "225*“ rate 0.28 0.28 0.28 0.28
population 0
Male marry-out rate (%) 0.043 0.043 0.043 0.043
Female marry-in rate (%) 0.19 0.19 0.19 0.19
Male marry-in rate (%) 0.043 0.043 0.043 0.043
Leave horn: 0)13tentron rate 42.0 42.0 420 42.0
0
Rate switching to
electricity from fuelwood N/A 70 70 95
for cooking (%)
Rate switching to
FuelWOOfi electricity from fuelwood N/A 35 35 65
consumptron for heating (%)
Rate switching to
electricity from fuelwood N/A 25 25 35
for raising pigs (%)
Thinning cycle (years) N/A N/A 1 1
. Age to start thrnnrng N / A N / A 10 10
Reforestation (years)
for fuelwood Thinning intensity (%) N/A N/A 3 3
Age to end thlnnrng N/A N/A 60 60
(years)

 

 

 

 

 

 

78

Table 4.2: Scenario evaluation results from 36 evaluators

 

Scenario A Scenario B Scenario C Scenario D

 

Number of evaluators 36 35 35 33
Average scores 4.41 5.55 5.53 6.33
Standard deviation 2.28 1.95 1.86 1.67

 

Note: the average scores of the four scenarios are signiﬁcantly different (F 3,128 = 8.33,
p < 0.001 from a one-way ANOVA test) and the mean score of Scenario D is
signiﬁcantly greater than the second biggest score of Scenario C (t = 1.78, p = 0.039, df
= 66 from a one-tail t test).

79

Figure 4.1 Study site — Wolong Nature Reserve

  

‘ ‘ ,3 . Elevation(m)
:' . -<2.000

1 2.000 - 3.000
' " 3000-4000
['3 4.000 - 5.000

. >5.ooo

Map of China Wolong Nature Reserve

80

Figure 4.2 Interactions among pandas, people, and policies

 

   
   

  

Pandas
(Habitat)

Kirrueno
A11119110

Distribution

 

 

 

 

 

 

U
('D
e :1. z a
0.3 People E. 2 CD
5‘ U)
.53.

Socioeconomics Others

 

 

 

 

 

 

Note: NFCP, Natural Forest Conservation Program; GTGP, Grain-to-Green Program

81

Figure 4.3 Three-tier architecture of the WOW

 

Presentation

 

Web browser

 

 

 

 

Web server

 

 

Business

 

 

 

Application server

_—————_—-‘

 

 

 

Data r ‘‘‘‘‘‘ 1

 

 

 

 

82

Figure 4.4 A typical work ﬂow for a new user in the WOW

 

Start a browser, e.g., lntemet
Explorer or Netscape

 

 

Go to the website of WOW:
http://panda.csis.msu.edu/wow/

 

 

[ Browse background and help pages ]

 

 

[Register as a new user]

1

L Log onto the system ]

 

 

 

[ Construct_a scenario *—

1

[ Simulate the scenario ]

Enter preference
weights of criteria
[ Evaluate the scenario )—

l

[ Log off the system ]

 

 

 

 

83

 

85:000. :0 wag—Sam .-
85:000. :0 m0:00m o

 

 

09$ ,3 “00:08::
“850: :0 0N5 0

 

 

8:38:06 3:03
5:00:00 803—03 0

 

 

ﬂ 8:090 £05005 0

 

 

ﬂ 5:03:05 “0003—02.: 0

 

 

8:995:00
“0003—0?— 6

 

 

0N6 8:038: 6
0.205030: :0 0:832 o

 

 

  

 

 

 

00:0:000 m0mm0mm<
00:00:05 .000: :0 :03: 85:08 00:00m
.2008 wﬁxmﬁ :0300: «tour—0:32

 

 

800:0.“

.880: E “00:08:: 2:50: 05 M0 0.5 0.: 32009
.2005 8083030 809:: 2:30:

 

#

 

 

300.8,: ~830:
E 000—: 000:: :0:00:00 “0003—04: 80:3 30:00.:—

1

 

6:08 :0::£.:m:0 05 “00020:”—

 

9

 

 

80200.: 82:33. 0: 30500—0 :0 8:095”—
_0:0:: :0:00=00 803—0.: :0 83:: 0::0:00m

 

 

d

 

 

£58m “00:00:80.. 9:855 :8: coca—0:“:
3:08 803—08 :0: :0:Sm0:0.:0m

 

 

F

 

 

0.00603— 28 5:5: :0: @0022: m0 850:8 oft
3:08 5:955:00 8020:”—

 

 

»

 

 

03:030.: :0 5:20me 60:80:00. .:0:00:U
”2:203:05 .«0 080: 260— dwathﬁ: .500: .atmm
3:08 0280:»: 8:330: 5:5:

 

 

 

80:20:05 6
«£210 a

 

 

as: use: 0:355 sass: .

F k

W 1
000.8 :0:00=00 “0003—03 _:::0:0m 6
830 @0022: 32.535 0

 

 

 

30:000. 20:00:02 0

f \

J

 

W

00% 305005 6

L

j
56:25 w::m0>.8: o

630% :53 50:0“: o

f

 

\

 

A02 dwm 6&3 3:03 330.53% 6
F L
N J
9830.5 6
5:538: £336 6
L
x 3
0:8 0:8: 0201— 0
0:0: 0w0=00 o

 

 

 

(

 

 

 

f 0:: gm ok

84

Figure 4.6 Algorithms of predicting the location of ﬁJelwood collection

 

Household with fuelwood
consumption, PFAs (potential
fuelwood collection areas)

 

 

 

 

   

Find the nearest PFA with
enough fuelwood stock and
assign it to the household

  
 
      

Fuelwood stock in
the PFA > Fuelwood
consumption

 

 
      
 

 

[ Get the extent of the PFA polygon ]

l

[ Randomly generate a point inside the extent I

 

 

No

  

 

Is the point inside
the PFA polygon?

 

 

[ Update fuelwood stock of the PFA ]

 

85

Figure 4.7 Algorithms of determining whether or not a household collects fuelwood

from natural forests

Household fuelwood
simulation starts

 

 

  
     
    

Does scenario
consider plantation
forests for fuelwood?

 
  

 

K Gets total fuelwood consumption fwc for all N \

households and fuelwood supply fws from harvesting

plantation forest stands. Calculates the percentage of
fuelwood from harvesting:

 

 

\ p=fws/fwc;i=l j

l

 

Households need

natural forests for
fuelwood. Find
locations for them.

 

 

Gets i‘h household and its
fuelwood consumption fwc,-

 

No

 

 

 

Generates a random number r
between 0 and l

 

 

V

No Need natural
forests for
fuelwood; find a

No

 

 

[h

location.

Ye

Move to next
ousehold: i= i+ 1

t

 

 

 

 

No need to cut natural]
forests for fuelwood J

 

86

]‘

Figure 4.8 Screenshots of the WOW illustrating (a) the layout of the main page with
four parts, (b) models, (c) simulation results, ((1) scenario management, and (e) GIS
layers in map, and (f) menus used for navigation into other pages

a

1 WOW - Wolong Online WebGIS for Pandas, People and Palmer - Windows lntemet Explorer

 

o v e r»; 2pm; mud-mm . v -, x 9 -

4' 2;- tWOW—wwﬁgmwmﬁkxpm‘ I. ' , a ' V'ane' .,*9>t'0'£ 2$

  

WOW: Pandas-People-Policies in Wolong Nature Reserve

 

Center for Svstels Integration and Sustainabilitv. lichiun State L'niversitv

 

 

 

 

 

 

 

l
{ Model Parameters Username:test a Q n g 0 “mm“ 21. L06;
; Demography IFuelwood [Reforestation [Simulation I Scenario: 29 2006 ii’ 7
‘ Away: mwcm: is Ilorry Age: ' 5 WM“.
l wanna: in 'mmyommum): inc ' ‘ , a
, ~ 7 ' - ' Nouse‘v -
1 rumour-new): lu- nu-mnmuuumyenu 7 (“S
i l' .‘9 ' r \ ,. 192 ‘ Freda“
‘ ' 7 . : Fm-o..$
‘ Luveuoanelmendonute(ﬁ).l20 4 FMUJHID
| r ‘ ' .. 1 Stan sundabon i ‘ 1" 9"““3”
l L A a
‘ mm A ...4...:.. "—1

 

l , ,
l
l ‘ _ warmer ’J-‘t 19v 0 15

l

< )
f N
_ Hat-ma S u m 94 m mi L? I
‘_E-7‘vm:‘vv' '1 ‘
l

eledﬂdly expense [;:§I‘3{’ QKWJQFIU

 

 

 

 

 

 

 

l I .51; My CriterialMy ScenariosIEvaluate Scenariosl
i ' lﬁ‘f/ swezoloow; Three Ielthsshouldsomlo 100%.

l “Rm Ecologlcd magniﬁo Eamon-ac Vldghl 50

9’" ‘ m ,7 , ,, W ,
l , ,w'” K ‘) Soda! “ﬁght in UpdateMndleni‘]
Lu" a w >5 " r "‘—"

l l ,2: 3244‘“ " '54:
1 ’r,ll(‘ ‘ _ LW l

l WWW

‘ mi, c l

as" ,__ ,

I ' ’ 4 e = l] I '1 ' 15 re r

Li"? .L .4 .1 V L \ n 1 a. p _3‘“"“.’?" " "' " ‘“ 1 a z ,

 

 

87

Figure 4.8 continued

 

Model Parameters

 

Demography I Fuelwood [Reforestation [Simulation
Average Births per Couple: l2—5— Ilarry Age: E—
Upper Birth Age: :4?“ Male Many Out Rate (%): W
Female Harry Out Rate (%): [6’55“ Male Bring Female in Rate (%): WIT—
Female Bring Nale in Rate (%): W College Rate (%): W—

leave Home Intention Rate (%): Bill _.___

 

 

[ Reset gsmmeteeto default [L Stansumulatgoﬂ

 

 

88

Figure 4.8 continued

 

Check to show results on affected habtat by habitat type; Uncheclt to show simulation
results of human population, households, fuelwood consumption and production, and

electricity expense, r Show other results

_ .v..__,.
v.- 17- -“l ULE

 

 

 

 

.—
_F‘

.—

_ Number of Households

_ F uelwood Supplied in m3
_ Eleetnub/ Expense in 1000 0

131?

"3
‘r‘

 

 

 

' '7 ..
‘ I I, , I I a r‘-‘ , 11 I, "a
| I I“ I ‘:l[l ‘8“ ‘_; : W
. “ ‘ ‘ 4'3:
.1'! «~"‘
' ‘11:. :;"- __'1_L‘:'~_
hit¥i--‘
All 4? l ' I 1 «WM—
‘ ' _ _ _ 1‘21""
395 1"“ new
I .'| u: _ "I p——-
‘1 " . __.tt.
4 1 4L L _L J

 

p ‘2 4“5 ‘3‘ ‘10‘ ‘12‘ 14‘ ‘15‘ ‘18‘ 20

 

 

 

89

Figure 4.8 continued

d

 

My Criterial My Scenariosl Evaluate Scenariosl
Scale: 0-1000/0; Three weights should sum to 100%.

 

Ecological Weight W Economic Weight 30.0 ' .

Social Weight W[ UpdateMyCriten'a ]

 

 

 

 

 

9O

Figure 4.8 continued

 

Username:test
Scenario: 29

E] 5 WM Natl"
[a Place Nam.
91 Household:
Roads
[:1 Potential FL
Ci Fuelwood S
Panda Hab
Boundary

ETEiidﬂWEh

 

 

 

 

 

1 0.5 0 1Mlles
-:—

 

91

 

Figure 4.8 continued

‘ WOW Wolong Om‘lne Webb's for Pandu, People and Policies - Windows Internal Explorer
.- A
90' Ohm IU‘;312:“u£‘L’b,l

a. a’k‘ OW-wmmwmmsucm

”'1‘ P'

l! - 'ane' «TM'O'II 3

Center {or Syltnl [melt-Hon and Sustainability, lichi‘ln Sine University

Introduction

 

loam

92

Figure 4.9 F uelwood (a) demand and supply and (b) electricity expense dynamics of

four scenarios

 

 

 

—9— Scenario A
f -9— Scenario C
I + Fuelwood Production

+ Scenario 8
—6- Scenario D

 

 

__T__

- _'T' _.

 

 

 

 

 

14~
7 lZi
5 i
l-i’ 10‘
E A I
'UM

= E84.
3 e

If: E

'E *6:
’3
la

.5 4i
‘2 i
3 §
1““ 2:
’ I
l 0 ——‘ ‘
‘ l
b

350

l 300
3

5: 250
8.

X;—
me 200
>560
.::>
@3150
loo
9.—
;E"100
:: 50
C:

<

0 L
1

 

93

 

 

Figure 4.10 Impacts of four scenarios on panda habitat of four types: (a) highly suitable,

(b) moderately suitable, (c) marginally suitable, and (d) unsuitable

 

 

 

 

 

 

 

a
[’49— Scenario A + ScenarioﬁB —9-Scenariog +Scenario3
A 160 i , __
22‘? 140
DD 0
if S 120 i
5% g 100‘
a: '8 80‘
c.— m 60 “'5’ 2" - Q
C A a q
8 lg ¢ 3 ' 3 e e s ' ‘
:2 .e 40 e a .
a 20 * - g e 9
0 ‘ . i _ - - - . . A - - - _ a _
l 5 10 15 20
Year
b
i—B—ScenarioA + ScemrioB -€— chnﬂfio Ci—A— ScenarioDl
500‘ 7- , k ~_ ,, ,-s_s..a-____.__._.,

2

Suitable Habitat (1000 m)
N
o

—s
O
O

 

 

 

Area of Affected Moderately

O

 

 

94

Figure 4.10 continued

—9— Scenario A + Scenario B —9— Scenario C + Scenario D

 

 

 

 

 

 

 

 

>‘ 16 ' , ”dun," ..___ﬂ.___ _
ﬁg) N; 14
g 8 12 "
E 3 10
Q)
3 E 8
8‘5 :g 6
155 =
: a 2 4
12 1% 2
c3 0 1
1 1 5 10 15 20
i Year
d
['43— Scenario A —X—_Scenario B —9_—‘Scenario_ C —¢:_§c__enario_ D
A 700 - r . .7 _. _ m a,“
l 5. 3
2‘5 600 ,
a I , _ .
B 3 500‘ ,2- ., 1,4:—
e - - . - . -'-~-~
8:: ,2 400 ‘
< a
is I 300 F c Q's-9:, ;~ 0
‘3 .
i3 1% 200 g 3 ‘3 3 3 e e e e
1 9' °
I a 100 e a = .
C: ‘ .‘. - a
i D O T “ ral‘slala alaﬁlarﬁ a—rAIGIE
i 1 5 10 15 20

95

 

CHAPTER 5

THINNING PLANTATION FORESTS FOR F UELWOOD:
SEARCH FOR SUSTAINABILITY

ABSTRACT
When wildlife and humans compete for natural resources and their needs conﬂict,
conservationists often blame human beings and tend to create nature reserves, where
resource-extractive activities are normally not allowed. Especially in developing
countries, human communities usually live a traditional subsistence lifestyle heavily
dependent on the extraction of natural resources, but possible sustainable resource-use
schemas are rarely explored. In the case of giant panda (Ailuropoda melanoleuca)
conservation in China, there is no exception. Human activities such as timber harvesting,
fuelwood collection, Chinese medicine gathering, agriculture, urbanization, and tourism
have led to habitat fragmentation and loss. Many of those activities, such as commercial
timber harvesting and fuelwood collection, have been completely or partially banned in
panda reserves. Requirements of such dramatic changes in human behaviors are often
deemed to fail in the long term because of economic and/or social reasons. This chapter
presents a case where a sustainable resource-use schema could be achieved with
appropriate policy design. I use Wolong Nature Reserve, the ﬂagship for giant panda
conservation in China, to demonstrate that its 967 ha of existing plantation forests can
sustainably provide approximately 1,000 local rural households with enough fuelwood,
the primary energy source for heating and cooking, if deliberate forest thinning plans
and conservation policy regulations were applied. This chapter may also shed some

light on the current energy crisis worldwide.

96

KEYWORDS: plantation forests, fuelwood supply and demand, thinning, sustainable

resource, giant panda, Wolong Nature Reserve, China

97

INTRODUCTION

Sustainability of conservation policy and practices is key to successful wildlife
conservation in the long run. In coupled human-nature systems (Liu et al., 2007a; Liu et
al., 2007b), with the increasing population growth and technological development,
human communities often overexploit natural resources such as forests on which
wildlife also depends and thus endanger the survival of wildlife populations. Though it
is arguable that conservation should be based on biocentralism or anthropocentn'sm
(Lowry, 1984; Brown, 2003; Redford and Sanjayan, 2003; Chan et al., 2007),
conservation practices considering human interests without impairing the environment
can often be more acceptable and lead to successes (Arambiza and Painter, 2006;
Kaimowitz and Sheil, 2007). Therefore, conservation policies and practices that create
ecologically sound, economically feasible, and socially amiable sustainable
prescriptions for human society are vital (Robinson, 2006; Chan et al., 2007). However,
in reality, such potential sustainable resource-use options are rarely explored
deliberately by conservation decision makers before creating protected areas and
banning extractive human activities within such regions (Salafsky et al., 2001). It may
be because of the political pressures from higher government administration and
international interest groups (Orr, 2003). It may also be because of the ignorance of
knowledge about sustainable natural resource uses (Boot and Gullison, 1995; Van
Rijsoort and Zhang, 2005).

Thus, this chapter aims to present a case study of searching for sustainable natural
resource uses in a protected area. This effort was assisted by WOW, a WebGIS-based

DSS. The study site is located in mountainous western China and cohabitated by

98

farmers and wildlife, including giant pandas. The farmers live in a traditional
subsistence lifestyle and traditionally cut trees for fuelwood from natural forests, which

often are suitable wildlife habitat (He et al., in review). A survey of fuelwood collection

. th
locations over the last three decades of the 20 century reveals that the farmers had to

select gradually more distant sites at increasingly higher elevations, and the majority of
tree felling occurred in areas of highly suitable and moderately suitable panda habitat

(He et al., in review). The annual amount of fuelwood consumed by each farmer

household ranged from 8 to 30 m3, depending on household size, age structure,

cropland area, and other socioeconomic conditions (An et al., 2001). From 1975 to 1998,

the human population and the number of households increased by 69% and 124%,

respectively, while fuelwood consumption in the reserve doubled from 5,000 m3 to

10,000 m3 (Liu et al., 1999c). Electricity, an ideal substitute of fuelwood, is expensive

and was mainly used for lightening and other household electronic appliances such as
radios and TVs before 2000 (see chapter 3). With subsidies from several conservation
programs (discussed later) started since 2000, households have begun to buy electricity
for cooking human food and heating houses in winter (see chapter 3). This switch did
not totally reduce the dependence of household energy need on fuelwood. A. sustainable
source of fuelwood for rural households in the reserve is necessary in the long run,
especially in the context of the current energy crisis worldwide (F argione et al., 2008;
Sedjo, 2008) and natural disasters such as the recent earthquake in southwest China

(Wang et al., 2008).

99

Considering the acreage (967 ha) of existing plantation forests in the reserve
(Table 5.1), we propose harvesting those forested stands through thinning as alternative
fuelwood sources. Most plantations are monocultures of Japanese larch (Larix
kaempferi (Lamb) Carr.) with a high tree density. These even-aged stands are not
suitable habitat for the panda because there is no understory bamboo and they are close
to roads and local residents, thus prone to human disturbances (Bearer, 2005). If these
areas were aimed for panda habitat in the future, the plantations would have to be
thinned and bamboo and/or other tree species intercropped to allow the adjustment of
forest stand structure in the long run. Besides this, there are several other beneﬁts of
collecting fuelwood in plantation forests. Obviously, it reduces the destruction of
natural forests. Because most of these plantation forests are close to human settlements
and natural forests are farther away, less time is needed for collecting fuelwood. As the
electricity price increases, plantation forests can also decrease the economic burden of
local residents, who can then spend the saved money for other uses. The availability of
fuelwood from local plantation forests is better than other energy sources, especially in
the event of natural disasters (e. g., earthquake).

Thinning is a natural process for forest stands responding to tree competition for
light and it takes place when intense competition from surrounding trees starts to slow
tree growth. The naturally annual thinning rate of a Japanese larch plantation varies
from 3.2% to 4.3% (Su et al., 2003). Proper artiﬁcial thinning can keep volume
productivity and even increase the timber yield and quality (Dong et al., 1999;
Archabald and Naughton-Treves, 2001). A strategy that the ﬁrst thinning occurs at 7—9

years old and the second 4—6 years later was introduced in southern China (Li et al.,

100

2000), while one in which the ﬁrst thinning with high intensity occurs at age 12—13
years and 5—6 years before the second thinning (with low intensity) in northern China
was also documented (Dong et al., 1999). If stands are managed for timber, 2—3
thinnings with long thinning cycles and high thinning intensities (e. g., 10—30%) are
often practiced (Li et al., 1992; Dong et al., 1999). Both tree stems and branches from
thinning could be used as a source of fuelwood. Studies on Japanese larch stand
biomass (Su et al., 2003; Zhang et al., 2005) reported that the ratio of branch biomass to
stem biomass ranges between 1:5 and 1:3.

Several questions are critical for conservation managers to answer in deciding
whether to adopt thinning of plantation forests to meet fuelwood needs of local
residents: Does it provide enough fuelwood for rural households? If so, when is it
available? and What is the best thinning schema, deﬁned by start thinning age, end
thinning age, thinning intensity, and thinning cycle? To answer these questions, we used
a systems model to simulate dynamics of households, estimate fuelwood consumption,
calculate fuelwood production from thinning, estimate the cost of electricity expense,
and predict the impact of fuelwood collection on panda habitat under different scenarios
over a period of 30 years from 1997 to 2026. Then we compared fuelwood consumption
and production under various scenarios to assess the sustainability of thinning
plantation forests for fuelwood in our study site, Wolong Nature Reserve.

METHODOLOGY

Study Area

Wolong Nature Reserve, created in 1963, is located in southwestern China and

comprises 2,000 km2 with elevations from 1,200 to 6,500 m. About 10% of China’s

101

wild pandas and 4,500 farmers in approximately 1,100 households cohabitate the
reserve (State Forestry Administration, 2005). Most households are located lower than
2,500 m in elevation. Pandas usually use areas with elevations from 1,500 m to 3,500 m
(Schaller etal., 1985). Panda’s staple food sources of arrow and umbrella bamboo
(Bashaniafangiana and F argesia rebusta) disappear in the areas with elevations higher
than the upper limit of 3,500 m. The other major component of panda habitat is forests
used as shelter. Mixed conifer and broadleaf forests with understory bamboo located in
the elevation range of 2,250—2,750 m are the most suitable habitat for giant pandas
(Schaller et al., 1985). On the other hand, most farmers also use these forests for their
subsistence economy and extract “free” forest resources through agricultural expansion,
road construction, house building, Chinese medicine gathering, and fuelwood
collection.

Since the establishment of the reserve, the local governments have tried to ban
fuelwood collection but rarely succeeded (He et al., in review). They have wanted the
residents to use more electricity, but the substitute electricity was low quality and too
expensive for most rural households. Since 2000, GTGP has started to revert cropland
to forests or bamboos through payments to households and NFCP has subsidized
households to monitor assigned forest parcels for illegal harvesting (Liu et al., 2008b).
Meanwhile, electricity quality and availability have been greatly improved due to a
national enhancement program of rural utility systems (Wolong Nature Reserve, 2005b).
NFCP and GTGP have explicitly forbidden any type of fuelwood collection. During the
beginning of these programs (2000—2003), few instances of illegal harvesting or

collection of branches were seen, partially because many households still had some

102

previous years. As a result, it was also observed that 60—80% of the energy needed for
cooking human food and about 30410% needed for heating houses in winter were
switched from fuelwood to electricity. Since 2004, as the leftover fuelwood was used up,
the gathering of dead tree branches has been seen and even allowed ofﬁcially (Chen et
al, unpublished).

Fuelwood is also important for the residents to keep their traditional lifestyle. One
of their major nutrition sources is smoke-dried pork, which also attracts tourists and
transfers their crops to cash. A sustainable source of fuelwood is urgently needed.
Fuelwood Consumption
Fuelwood consumption by households consists of three components: for cooking human
food, for heating houses in winter, and for cooking pig fodder. The volume of each
component (in m3 ) is a function of economic and demographic attributes of the
household as described below (An et al., 2001).

For cooking human food = 0.467 X Household size + 0.703
For heating = 8.63 if household has one or more senior members; 7.23 if not
For cooking pig fodder = 24.856 x corn land area (in mu, 1 ha = 15 mu) / 15.0

After NF CP and GTGP were started, households switched a portion of fuelwood
consumption to electricity. We observed switches of approximately 70%, 35%, and
25% for each component mentioned above, respectively. After 2000, we considered this
reduction of fuelwood demand. Another change after 2000 was that an average of 45%
of each household’s cropland was converted to tree and/or bamboo plantations. Thus,
the acreage of corn land was reduced proportionally to calculate fuelwood needs for

cooking pig fodder after 2000.

103

Fuelwood Production from Thinning Plantation Forests
A total of 487 ha of areas previously harvested for commercial timber have been
reforested since 1975 and 480 ha of cropland has been afforested since 1986 (Wolong
Nature Reserve, 2005b, see Table 5.1). The majority of those plantation forests was
monocultural and designed for timber harvesting and/or ecological services (e. g.,
erosion control). Less than 5% of cropland converted in 2000 was planted with species
(e. g., Longpeduncled alder, Table 5.1) suitable for fuelwood needs of local residents.
The most common species planted is Japanese larch, which is insect and disease
resistant (He, 1991) and fast growing (Yan and Liang, 1993) even on a very
nutrient-deﬁcient site (Kochenderfer et al., 1995). Japanese larch, though not the most
favored species for ﬁJelwood, has been consumed by households (He et al., in review).
Though Japanese larch has been present in Sichuan since the 19703, little is known
about its growth in plantations in Wolong or areas nearby. Only a few scattered records
were found in the literature (Li, 1982; Yang, 1989; The Editorial Board of Forests in
Sichuan, 1992; Yan and Liang, 1993). We choose the growth model developed by Ma
et al. (1998) in Liaoning Province, northern China, where Japanese larch was ﬁrst
brought into China in the 19205 and was believed to grow more slowly than in Wolong
(Yan and Liang, 1993). The model estimates the volumes as a function of site index (SI),
density, and age of forest stands. It is in the form of

0.3761

where V, is volume (m3/ha) and N, is density (trees/ha) at age t (years). We used the

only existing forest stand records from Wolong and nearby areas to validate these

104

 

where V, is volume (m3/ha) and N, is density (trees/ha) at age t (years). We used the

only existing forest stand records from Wolong and nearby areas to validate these
models. Table 5.2 shows that Japanese larch plantations in Miyaluo (located in Li
County, north of Wolong) followed the growth trend of their counterparts in Liaoning
with an 51 of 20. The model provides conservative estimations of the growth of
Japanese larch plantations in Wolong, an environment ecologically similar to Miyaluo.
Fuelwood production from stems could be estimated from the growth model discussed
above.

In summary, we ﬁxed the SI of Japanese larch plantation stands as 20, and used
equation 5.2 to predict volume production given stand ages and densities. For initial

values of parameters of stands planted in different years, see Table 5.3.

[5.2]
where V, = volume production (m3/ha), SI = 20, N, = stand density (trees/ha), and t =

stand age (years). The average ratio 1:4 of branch biomass to stem biomass was used to

calculate fuelwood production of tree branches from thinning. The fuelwood production

FWti from a plantation forest stand i at the year t when thinning occurs is calculated as
FW,, = f(Sl, Nu, t)/N,, x (N,, x T1,) x A, = f(SI, Nu, t) x TI, x A, , [5.3]
where TI, is thinning intensity of stand i, A, is area of stand i (ha), S1 is ﬁxed as 20, and

Nti is stand density of stand i at the year t (trees/ha).

105

 

by perturbing each major parameter by a certain magnitude, and calculating the
sensitivity index (J ergensen, 1986) as

Sx = (dx/x)/(dP/P) , [5.4]
where P is the value of the independent variable, dP is the value for a small change of P,
x is the value of the dependent variable, and dx is the corresponding change in x in
response to the change in P. We assessed the model by increasing each input
parameter—SI, initial density of forest stand, and the branch biomass to stem biomass
ratio—by 10% and calculating the sensitivity index for each year during simulations.
The minimum and maximum sensitivity indexes for each parameter are reported.

A thinning scenario can be deﬁned by thinning cycles, thinning intensity, ages of
thinning, and mature stand ages (when harvested). To examine how those parameters
affect the fuelwood production, we chose three groups of scenarios with different levels
of fuelwood demand (i.e., less, status quo, and more). We keep mature stand age at 60
years unchanged, which allows all the plantation forest stands to remain unharvested in
a simulation of 30 years. Thinning cycle varies from 1 to 3 years; thinning intensity
from 3 to 15%; and age to start thinning from 10 to 14 years. We designed the scenarios
in a way that if the thinning cycle increases by 1 year, thinning start age is delayed by 2
years and thinning intensity doubles. Three groups of the thinning scenarios with
different combinations of parameters are listed in Table 5.4. The status quo scenarios
reﬂect the situation of household fuelwood consumption reﬂected in the empirical data:
with the implementation of NFCP and GTGP, the rates of switching to electricity ﬁom
fuelwood for cooking human food, heating, and cooking pig fodder are approximately

70%, 35%, and 25%, respectively. Scenarios with more fuelwood assume that without

106

for cooking human food, heating, and cooking pig fodder are approximately 70%, 35%,
and 25%, respectively. Scenarios with more fuelwood assume that without NF CP and
GTGP in place households do not replace any portion of fuelwood consumption.
Scenarios with less fuelwood assume that if NFCP and GTGP increase their
investments households may meet their energy needs by switching to electricity. In this
case, the three switch rates from fuelwood for cooking human food, heating, and
cooking pig fodder are 90%, 65%, and 35%, respectively.

Implementation

Models for household fuelwood consumption and production from thinning plantation
forests were incorporated and implemented in WOW. WOW includes models to
simulate the dynamics of rural human population at the individual and household levels
(Human population dynamics model), calculate fuelwood consumption based on
demographic and social factors of households (Fuelwood consumption model), estimate
fuelwood production if thinning plantation forests (Reforestation for fuelwood model),
and predict the locations of fuelwood collection (F uelwood site distribution model) and
the costs of using electricity for each household (Economic impact of fuelwood
collection model). Then the system evaluates the dynamic impacts of fuelwood
collection on panda habitats at the reserve landscape level given different policy
scenarios (Habitat impact assessment model). Based on the multicriteria
decision-making (MCDM) methodology (Malczewski, 1999), a policy scenario
evaluation component (Multi-criteria decision making model) was also developed to
help decision makers assess and compare different policy scenarios and then choose the

best based on their own ecological, economic, and social preferences. For details on

107

design and implementation of those models and the system, see chapter 4. Because most
models are stochastic, each scenario was simulated with 20 replicates to examine the
fuelwood balance of households and the impact of fuelwood consumption on panda
habitat during a simulation of 30 years. The averages of measurements (e. g., fuelwood
demand and areas of impacted suitable habitat) were reported.
RESUTLS
Sensitivity Analysis
Table 5.5 lists the minimum and maximum sensitivity indexes of the three model
parameters against model output fuelwood production during a 30-year simulation. Site
index is the most sensitive parameter, followed by initial stand density and branch
biomass to stem biomass ratio.
Scenario Simulations
The results show that thinning plantation forests can satisfy a substantial portion of
ﬁrelwood demand (Figures 5.1—5.3) and lessen the impact on suitable panda habitat
(Figures 5.4—5.6). However, the fuelwood-balanced year, when plantation forests
produce enough fuelwood to offset household demands so that no natural forest cuts are
needed and the size of affected suitable habitat is zero, is different under different
scenarios. Most status quo scenarios and all the scenarios with less fuelwood can reach
the balance during the 30-year simulation.

None of the scenarios with more fuelwood (Bl-B9) reaches a fuelwood-balanced
year during the 30-year simulation length (Figures 5.1 and 5.4). However, during some
years (e. g., the harvesting years after year 17 for Scenarios B5 and B6 (Figure 5.1b),

and after year 20 for Scenarios B7, B8, and B9 (Figure 5.1c)), fuelwood production is

108

greater than demand. Fuelwood demands of the status quo scenarios are reduced by
40% after year 5. When there is no thinning rotation (i.e., thinning cycle = 1 year), the
fuelwood—balanced years of Scenarios C1, C2, and C3 are years 28, 16, and 15,
respectively (Figures 5.2a and 5.5a). When 2 or 3 years of thinning rotations are applied,
thinning plantation forests during the harvesting years after year 6 (for C4, C5, and C6,
see Figures 5.2b and 5.5b; for C7, C8, and C9, see Figures 5.2c and 5.5c) can produce
more fuelwood than the reduced fuelwood demands. The extra fuelwood from the
harvesting years may be enough to offset the fuelwood demands of the following
non-harvesting years before the next harvesting and thus no suitable habitat is affected
(Figure 5.5). The fuelwood-balanced years of some scenarios come 10 or more years
later (years 18 and 16 for Scenarios C5 and C6, respectively; year 18 for Scenarios C8
and C9), while Scenarios C4 and C7 can not yield the balanced years during the 30-year
simulation. The scenarios with less fuelwood reach a fuelwood-balanced year earlier
than the status quo scenarios: year 9, 6, and 6 for Scenarios D1, D2, and D3,
respectively, with no thinning rotations (Figures 5.33 and 5.6a); year 10, 6, and 6 for
Scenarios D4, D5, and D6, respectively, with a 2-year thinning rotation (Figures 5.3b
and 5.6b); and year 12, 6, and 6 for Scenarios D7, D8, and D9, respectively, with a
3-year thinning rotation (Figures 5.3c and 5.6c).

Of the status quo scenarios, the earliest one to reach the balanced year is Scenario
C3 (year 15). Most of the scenarios with less fuelwood reach the balance in year 6.
Since there is still extra unconsumed fuelwood in the scenarios with less fuelwood, the
switch rates from fuelwood to electricity were decreased in order to reduce the

electricity costs of households. The switch rates used in those scenarios for cooking

109

human food, heating, and cooking pig fodder are 90%, 65%, 35%, respectively. After
several tests of combinations of lower values of these three rates in WOW, we found
that using 80%, 30%, and 30% rates, respectively, and the same thinning schema as in
Scenario D3, the fuelwood-balanced year of the newly created scenario (named CD3)
would be year 15 and the electricity cost would be reduced to the status quo level.
Because the status quo switch rates (70%, 35%, and 25%) are close to these, NF CP and
GTGP may not need to invest too much in order to achieve the higher rates as in the
scenarios with less fuelwood. This effort also demonstrates a way to use WOW to find a
better decision alternative.

DISCUSSIONS

We demonstrated the feasibility of thinning existing plantation forests in Wolong
Nature Reserve to meet the fuelwood demands of approximately 1,000 rural households.
The earliest fuelwood-balanced year of the status quo scenario (C3), which does not
require rural households to collect fuelwood from natural forests, is year 15 of
simulation, and scenarios with less fuelwood can reach a fuelwood-balanced year as
early as year 6 and still produce surplus fuelwood. An optimal scenario (CD3) for
which rural households spend less on electricity than in scenarios with less fuelwood
was found through WOW. However, it is worth discussing several limitations of this
analysis.

We assumed that energy need and consumption structure do not change over time
and household fuelwood demands are still functions of social and economic
characteristics of households. However, as the difficulty of collecting fuelwood

increases, because of stricter law enforcement and less availability of ﬁrelwood nearby,

110

and the price of electricity also increases, households will likely reduce the demand by
possibly increasing use efﬁciency, thereby sacriﬁcing human welfare; for example, in
winter people may start heating later and stop heating earlier. If this is the case, our
study overestimates the fuelwood demand and actually increases the feasibility of
thinning existing plantation forests because less forest acreage or less intensity of
thinning is needed to meet the lower demand. On the other hand, there may be more
households and households may be richer. Both may require more fuelwood. Our
models have considered the population dynamics during the 30-year simulations and the
latter deserves a further consideration.

In the status quo scenarios and the scenarios with less fuelwood, we assume that
NF CP and GTGP will continue and/or increase their investments during the simulation
periods so that households would maintain current fuelwood-electricity switching rates
or consume more electricity. This increases the economic burden of the governments.
One way to alleviate the burden is to use non-forested land suitable for forests and
shrub land suitable for fuelwood. Plots close to human settlements in such land,
potentially not suitable for panda habitat, can be chosen to plant trees for fuelwood.
Harvesting forests close to human settlements have minimal impacts on panda habitat as
those areas are rarely used by pandas (Bearer et al., 2008). The other way may include
increasing fuelwood productivity through better management of forest stands and
species selection (Bisong et al., 2007; Kang et al., 2007; Kumar, 2008).

Our results reﬂect the general growth trends of the plantations of Japanese larches.
To be more realistic, data on the growth of local forest stands should be collected to

verify and/or adjust the models adopted from the limited literature. To implement the

111

ideal of thinning plantation forests for fuelwood, a strategy similar to NF CP may be
adopted. Assigning plantation forest parcels to households for fuelwood and monitoring
illegal harvesting should be considered. Operational and administrative costs of
thinning forest stands, harvesting and monitoring fuelwood collection, and distributing
fuelwood to households, were not main themes in this analysis. An economic feasibility
analysis considering these costs and beneﬁts, such as reducing household expenditures
of buying electricity, should be conducted while evaluating and choosing optimal
thinning scenarios and before implementing the selected scenarios.

CONCLUSIONS AND CONSERVATION IMPLICATIONS

Using systems modeling and scenario simulations, this chapter demonstrates the
potential that harvesting 967 ha of plantation forests can provide a substantial portion of
the energy needs of local rural households without cutting natural forests in Wolong
Nature Reserve. Meanwhile, this may keep the traditional fuelwood-dependent lifestyle
of minority ethnic groups in the reserve. Most of the status quo scenarios and the
scenarios with less fuelwood can reach the fuelwood-balanced year between year 6 and
year 28 of simulation. Because thinning can maintain the health of forest stands and
provide an opportunity to adjust the monocultural structure of plantation forests (e. g.,
interplanting bamboo) toward suitable panda habitat, we encourage reserve managers to
consider this idea and incorporate it with on-going conservation programs such as
NFCP and GTGP as soon as possible.

It is worth noting that the intention of the study is not to replace electricity with
fuelwood but to diversify energy sources of rural households. Efforts to improve quality

and decrease price of electricity should continue and complement the idea of harvesting

112

plantation forests for fuelwood. This study builds a useful foundation and provides
reserve managers with scientiﬁc information and tools to explore such a proposal of
sustainable natural resource use that beneﬁts both local people and pandas in the reserve.
The promising results also have important implications for the current global energy
crisis. Plantation forests, one of many renewable biofuel resources, may continue to be
used as a traditional and sustainable energy source, especially in those remote and rural

regions with long-lastin g conﬂicts of human development and wildlife conservation.

113

Table 5.1: Areas of plantations since the establishment of Wolong Nature Reserve
Species planted % of cropland
Japanese larch (Larix kaempferi Corn),

1 975_] 977 487.06 C hrnese larch (Larrrpotammr Batal. ), (cu tover areas)
masters larch (Larzx mastersrana),

Chinese spruce (Picea asperata Mast.)

Time Area (ha)

 

 

 

 

 

 

 

 

 

1986 1 13.00 Japanese larch 15
Japanese larch, Chinese ﬁr
2000 266'60 (C unninghamia lanceolata (Lamb. ) 36
Hook), maple (Acer L.), longpeduncled
2001 66.67 alder (Alnus cremastogyne Burkill), 9
a Japanese cedar (Cryptomeriajaponica
2002 81-83 (L. f) D.Don), water ﬁr (Metasequoia 1 1
I I ' I
2003 34.00 gyp ostrobordes Hu et C zeng) 5
Subtotalb 562.00 76
N/A

 

Grand Total 1049.06

 

a In 2002, cropland converted for bamboo plantations was not counted for fuelwood

roduction
Sum of cropland converted since 1986

Note: N/A = not applicable.
Source: compiled from government documents (in Chinese) (Wolong Nature Reserve,

2005b)

114

 

 

 

Table 5.2: Comparisons of characteristics of Japanese larch plantations in the model and g
Miyaluo
Source Age (years) DBH (cm) ”a?“ (3216);}; SI 1:31:70?
Modela 20 12.5 13.78 2,006 20 192.00
Miyaluo ab 18 14.6 10.52 2,080 NA/ 183.6
Miyaluo 6C 20 14.7 13.6 N/A NA/ 224.3

 

a From Ma et al. (1998)
b From Yang (1989)

c From Yan and Liang (1993)
Notes: DBH = diameter at breast height; 81 = site index, N/A = not available.

115

 

Table 5.3: Initial values ofpararneters of stands planted at different times

 

Year planted Site 1ndcx Inrtral densrty Age in 2005 Year of ﬁrst

 

(S1) (trees/ha) (years) thinning
1975-1977 20 1,288a 28 2005
1986 20 2,0068 19 2005
2000 20 3,976b 9b 2006
2001 20 3,976" 3" 2007
2003 20 3,976" 6" 2009

 

a Derived from Wang and colleagues’ model (1996).
b Data from Moy (2005).

116

 

Table 5.4: Three groups of thinning scenarios

 

 

More Status Less Thinning Asirio Thinning Age to. end
fuelwood quo fuelwood (52:11:) thinning 1nt(e)23)1ty $316123?
(yeaIS)
Bl C1 D1 1 10 3 60
I32 C2 D2 1 l 0 4 60
B3 C3 D3 1 10 5 60
B4 C4 D4 2 12 6 60
BS C5 D5 2 1 2 8 60
B6 C6 D6 2 l 2 1 0 60
B7 C7 D7 3 14 9 60
B8 C8 D8 3 ‘l 4 1 2 60
B9 C9 D9 3 14 1 5 60

 

Note: The rates of switching to electricity ﬁ'om fuelwood for cooking, heating, and
cooking pig fodder are as follows: 0% for scenarios 81—89; 70%, 35%, and 25%,
respectively, for scenarios C1—C9; 90%, 65%, and 35%, respectively, for scenarios
Dl—D9.

117

 

Table 5.5: Sensitivity test results for 30 years of simulations

 

 

 

+10% Sensitivity Index
Parameter Default value rtu b t‘
p e r a ron Minimum Maximum
Site Index (SI) 20 22 1.490 1.493
Initial stand 1288, 2006, 1416.8, 2206.6,
density 39763 3614.52‘ 0.361 0.492
Bram“ b‘omass ‘0 0.25 0.275 0.198 0.202

stem biomass ratio

 

a The ﬁrst two numbers are the initial densities of stands planted in 1975—1977 and
1986, respectively; the last number is for total stands planted in 2000, 2001, and 2003
(see Table 2 for their default values).

118

 

Figure 5.1 Comparisons of fuelwood demand and production of the worse scenarios (a)
81-83, (h) 84-86, and (C) 87-89. Black ﬁlled squares represent fuelwood demand and

non-ﬁlled shapes are for fuelwood production. Fuelwood demands are the same in the

three ﬁgures (a, b, and c)

‘ + Fuelwood Demand + Scenario Bl —°— Scenario 32 + Scenario B3

20 , ; . _; _ ‘ ” . ' ;LL:" _- if.
I
l

 

 

C 1
g"Elsi
Es '
.52
15:10 5
113.2 '
183
2% 5'
we 1
in.
l 0'
1 5 10 15 20 25 30'

 

119

Figure 5.] continued

1
25-

20'

IS

Fuelwood Demand and

Production (1000 m3)

5.
0.

120

' 1' l l T ‘— ‘T—T"—T' 1

Year

1

2

ll

‘ Ill
. a i i i f ‘ “ “
‘° 1 t /““‘VVVVV WW \I‘

‘ + Fuelwood Demand —9— Scenario B4 -9- Scenario BS + Scenario B6 J

 

i l

,0

1 ll 0
0 01
OT _

25

.11

I‘I

I/ \I‘ 0‘

30

 

 

Figure 5.] continued

i + Fuelwood Demand ‘9' Scenario B7 + Scenario B8 + Scenario 39

D
1 . __ r ___ |
‘

 

   

 

EMA 30- ﬂ ’.
E <5 25 1 ,i Ii f
i” i i
. 5 E 15‘ 4 ; : ‘\§: //A\ I/A\
? “’2 w '; i ”\ \ /"\ /' e , \
I E E 5 vvvv \ I \ “x ..
I 1 5 10 15 20 25 30

Year

121

 

Figure 5.2 Comparisons of fuelwood demand and production of the status quo scenarios
(a) C1-C3, (b) C4-C6, and (c) C7-C9. Black ﬁlled squares represent fuelwood demand

and non-ﬁlled shapes are for fuelwood production. Fuelwood demands are the same in

the three ﬁgures (a, b, and c)

—' m7, ‘ _.. g 4 ”n ‘

 

‘ i+Fuelwood Demand +Scenari0 Cl +Scenario C2 “ﬁr—Scenario C3'
\ ‘. ,,._. a ,_ # __ J
«am/K ! 9
1 5 E '
5 g g ‘01 . g __=_:______,--_
i E C I ”r 2 : _
Q) C .
a .2 - : ;
”£35k.
i E E i '
. a) m
:3
I LL- Ot 7 " ‘ ' I ' ' i r r I '7 ’r i T ' I ' ’T’ 7 r‘ T I V r’ﬁ‘ﬁ’MT
; l 5 10 15 20 25 30
Year

 

122

Figure 5.2 continued

 

 

 

b

i V+Fuelwood Demand +Scenario C4 + Scenario C5 +Scemrio C6]

25. _,7,_', , _'_
2m; 20 '4‘ i l! ‘
«2e .5‘ /‘ /‘/‘\ i
E 3 . /\g * \/‘\ /:
5 .53 10V /\i Hg A
WWW/WM" V V/ W AV V

 

20 25 3O

 

Year

123

Figure 5.2 continued

 

V i + Fuelwood Demand ‘9‘ Scenario C7 + Scenario C8 + Scemrio C9 V

 

 

 

 

. 35; __
V 301 ‘;
VEME 25*
la 52: J V
.E 3 .
523 _§ 15:» ‘ ,
:‘3 ‘5 L. A 2 /A\ ,,
3 E '01 /=\ 4&V M *
3 == \/\ / /\
: 0 L, . “a .
I 5 lo 15 20 25 30

124

 

 

Figure 5.3 Comparisons of fuelwood demand and production of the better scenarios (3)
D1 -D3, (b) D4-D6, and (c) D7-D9. Black ﬁlled squares represent fuelwood demand and
non-ﬁlled shapes are for fuelwood production. Fuelwood demands are the same in the

three figures (a, b, and c)

 

a
‘ [+Fuelwood Demand +Scenario DI +Scenario D2 +Scenario D3J
V is I , ~+ a +5
V s
V A 9
.0 r
. g ”a V
‘2 g 10 - 2 , 2 g a
. EU 2 - g 2 S ‘
I E V . 2 : ‘
i Q) C o
i D .2 ' . " ‘
'0 w . . e ’
V 0 g 5 1 ‘ - .
P "U ‘ . e c ' " -—'-"' — -
3 e ‘ C ' - . 3 ._—-..'
Q) Q— : 5 ' -
I :3
Li-
I
0 i V r I r V V r T V T ﬂ” 1 V r V 1 I T I T—‘Yﬁ

 

 

125

Figure 5.3 continued

 

 

 

 

 

b
1— — g _ _ 7 _ m— — ”—ﬁF j”;—
1+Fuelwood Demand +Scenario D4 +Scenario D5 +Scenar10 D61
251 " if ' 74— " 1*“ -—~
Em; 201 1 '1 .1. ‘
E. E151/1111111111\1
a 3 1
5&1 11111
1111,1111 it +
g =- 1
“‘ 0
I 5 10 I5 20 25
Year

 

 

 

 

126

Figure 5.3 continued

 

 

 

 

 

C
T’- 7 H 7, —— ' "F :______4_ "— '— _—_—— __ ___4_7____ﬁ
+ Fuelwood Demand + Scenario D7 -°— Scenario D8 '13- Scenario D9]|l
35 1 I if '7' iAn—i ,4 "* ,44_* 44‘" . ___' -

A 30* . 1 111
Ema A 1 {1.1. 1
:1 § :3; /1\ 11 /1\ 1
a C . r 1
as .13 Is; \1 /\ 1
\«A\\A ‘

4 W1 1 1 1

B l— 54 ~

,V VI: , Vﬂﬂ
15 20 25 30

 

Year

 

127

 

 

Figure 5.4 Different impacts of thinning plantation forests on highly and moderately

suitable panda habitat of the more fuelwood scenarios (3) Bl-B3, (b) B4-B6, and (c)

 

 

 

 

 

87-89
a
1_ -9_—_ScenarioBl_l fe—s'cénzina'sz' als'éenano BX]
500 ~ —~ -- —— — — —.-_ - ___________ .444, 44
NE
1: o 400
2 8
I 3 c,
' ‘2: § 300
1.. :5
O «a
g E 200 ‘
< a . l . - A n A ‘ - - A . . i A
'5; 100
0 I ﬂ 1 1 T T f r V T T 1 I 1
1 5 10 15 20 25 30

 

 

128

Figure 5.4 continued

1 ~1— Scenario 11 :11 ' +1411 1
7m 4 ,, . __ _ _ ___ __ _ _ ._________ _____ ._

 

 

Area of Affected
Suitable Habitat (1000 rr21)

 

 

 

 

130

 

 

 

 

Figure 5.5 Different impacts of thinning plantation forests on highly and moderately

suitable panda habitat of the status quo scenarios (a) C 1 -C3, (b) C4-C6, and (c) C7-C9

 

1 +Scenario c1 ,

4% Scenario c2 ' ' 33921991 _e

L)!
O
O

l

J}.
O
O

(1.)
O
O

 

N
O
C

Area of Affected
Suitable Habitat (1000 m2)

O
O

 

 

 

131

 

Figure 5.5 continued

 

1 —9— Scenario C4 1-9- Scenario C 5 _ +Scenario_C_b_ 1

 

 

 

500' “ ‘ .-.__ ,4

NE I
1% $400 1
10 v
12 5300
'1... E
o «a
1“ :3, ° 4
: -5100* I n 1. I
’ m " AAA 1 - '
O ‘ - '4 - ’4' - '5 Al}
1 25 3O

1 l 5 10 15 20

132

 

Figure 5.5 continued

Area of Affected
Suitable Habitat (1000 m2)

 

500

400 ~«

300*

200‘

 

 

-9— Scenario C7

 

-9— Scenario C8

Stu—w

 

l 5
Year

133

3.— a!

>

 

20

—A— Scenario C9

 

 

 

Figure 5.6 Different impacts of thinning plantation forests on highly and moderately
suitable panda habitat of the less fuelwood scenarios (a) D1-D3, (b) D4-D6, and (c)

D7-D9

=. __ -El- Scenario DrﬂjﬁgScen-ario D2 —A— Scenario D3 J

.b
8

7

A rea of Affected
§

Suitable Habitat (1000 Iii)
N
8

§

 

 

 

 

Year

134

 

 

Figure 5.6 continued

 

 

i ‘B—ScenarioD4
A500‘
6‘5 {-1
g E400 .
its? 3
:5 E300
“5 §
I «a 33200 ' 9
1 8 2 u
< g c 9 v
.2100
5 A A U i
I
0 I'\ I. '3 ';¢. J ;‘
l 5 10

—9— Scenario D5 _ -ﬁ-Scenario D6

.............
......

135

AAAAAAAA
O O C O .

 

 

Figure 5.6 continued

w A
8 8

§

Area ofAffected
Suitable Habitat (1000 m2)

 

é

 

O

 

i—B—Scenariom :-9- Scenario D85 __ iii—Scenano D9 _
500 ______ "H - -

 

136

 

 

 

CHAPTER 6

SYNTHESIS AND CONCLUSIONS

The challenge to understand the dynamics of coupled human-nature systems across time
and space for balancing human energy needs and wildlife conservation can be tackled
with assistance from decision support systems (DSSs). In this dissertation, a Web-based
DSS named WOW was developed to analyze the interactions among human society,
policy interventions, and wildlife habitat for Wolong Nature Reserve for giant pandas in
China. With this system, various geographically dispersed stakeholders can simulate the
dynamics of human population and fuelwood needs, predict the locations of fuelwood
collection, and then assess the impacts of fuelwood collection on the human community
and panda habitat. Different policy scenarios deﬁned by the input parameters of models
in WOW can be constructed and evaluated based on the ecological, economic, and
social preferences of stakeholders.

The dissertation starts from an analysis of the spatial and temporal trends of

fuelwood collection in the last three decades of the 20th century (19705, 19805 and

19905). Two hundred rural households were interviewed for their locations of fuelwood
collection sites during the three decades and other relevant ecological, economic, social,
and demographic data. The surveys showed that households tended to use previous
fuelwood sites or nearby areas for fuelwood and tens of adult males from a number of
households collected fuelwood collectively in order to increase efﬁciency and minimize
risks in the topographically challenging high mountains. It was also found that

fuelwood collection sites were moving toward higher elevations that are more distant

137

 

and closer to highly suitable panda habitat. Consequently, fuelwood collectors were
traveling longer distances to physically challenging areas, in our case, to areas of
high-quality panda habitat. F uelwood collection 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 the
regulations on fuelwood collection. Furthermore, this analysis reinforces the importance
of addressing the needs of local communities for the success of conservation practices.

Recognizing the devastating impacts of human activities on ecosystems and human
society, the Chinese government has been implementing several conservation policies
in the reserve since 2000. These policies have direct and indirect effects on households.
The dissertation continued to use household surveys to show that the effects of NFCP
and GTGP on rural households’ switch from fuelwood to electricity varied for the three
fuelwood consumption purposes, of which the one for cooking human food was the
highest (66.92%) followed by the one for heating (37.05%) and the one for cooking pig
fodder (22.62%). The energy dependence of households on fuelwood was estimated
with a reduction of approximately 36%. This study implied that fuelwood as an energy
source for rural households may co-exist with other sources such as electricity for the
long term. The results also suggested the challenge of ameliorating the environmental
impacts of human activities through programs such as NF CP and GTGP and the
necessity of searching for other sustainable natural resource uses.

Results from these two studies were then incorporated into a fuelwood

consumption model and a fuelwood site distribution model in WOW. Other models in

138

 

 

WOW include human dynamics, reforestation for fuelwood, economic impact of
fuelwood collection, habitat impact assessment, and multicriteria decision making
(MCDM). WOW is accessible at http://panda.csis.msu.edu/wow. As a demonstration of
the capability of the system, four scenarios were designed to show the effects of NFCP
and GTGP and thinning plantation forest stands for fuelwood; human population
parameters were set the same for all scenarios. The implementation of NFCP and GTGP
may reduce approximately 30% of the fuelwood consumption by households but
households still need to cut natural forests to meet the remaining 70% of their fuelwood
demand (Scenario B). Thinning plantation forests could provide a substantial portion of
fuelwood needs: Scenario C may reduce household fuelwood needs by 50 %.
Presumably, if the switch rates from fuelwood for cooking human food, heating, and
cooking pig fodder are 90%, 65%, and 35%, respectively, the areas of affected habitat
of all types are gradually decreased to zero after the simulation year 10 (Scenario D).
Thirty-six out of 76 registered users from eight countries in four continents have
evaluated these four scenarios using the MCDM component in WOW. The preliminary
results showed that the thinning scenario (Scenario D) was preferred by the stakeholders.
Besides the evaluations, more than 50 new scenarios have been constructed and
simulated in WOW after being made public through the listservs of three
conservation-related organizations in early October 2008. It is expected that more
scenarios will be created and the ranks of existing evaluated scenarios will be changed.
As implied from the aforementioned studies, fuelwood will likely exist as an
important energy source of rural households in the long term and it is worth a further

search for an optimal schema of thinning plantation forests for fuelwood in WOW. The

139

 

 

As implied from the aforementioned studies, fuelwood will likely exist as an
important energy source of rural households in the long term and it is worth a further
search for an optimal schema of thinning plantation forests for fuelwood in WOW. The
earliest fuelwood-balanced year of the status quo scenario (C3) is year 15 of simulation,
and the scenarios with less fuelwood can reach balance as early as year 6 and still
produce surplus fuelwood. An optimal scenario (CD3) in which rural households spend
less on electricity than in the scenarios with less fuelwood but can still reach fuelwood
balance at year 15 of simulation was found through WOW. This exercise indicated the
potential that thinning plantation forests can satisfy a substantial portion of fuelwood
demand and lessen the impact on suitable panda habitat. This effort also demonstrated a
way to use WOW to ﬁnd a better decision alternative.

WOW lays a good foundation for analyzing the dynamics of coupled
human-nature systems and evaluating policy scenarios. It can also be used as a
communication and educational tool for increasing awareness of conservation from the
public. However, as most similar systems, it has its limitations. Based on feedback from
users, future efforts with WOW may focus on reﬁning the models and improving
usability and performance toward an operational and collaborative system.

There is no compelling evidence that the current global energy crisis may
disappear or slow down in the near future. Therefore, the methodology presented in this
dissertation also has important and broad implications in the search of sustainable
solutions balancing human energy needs and wildlife conservation worldwide,
especially in those remote and rural regions with long-lasting conﬂicts of human

development and biodiversity conservation.

140

BIBLIOGRAPHY

Adinarayana, J ., Laurenson, M., Ninomiya, S., 2006, Web-based decision support
system for rural land use planning - WebLUP - a prototype, Agricultural
Engineering International 8:Manuscript IT 05 005

Aigner, P. A., Block, W. M., Morrison, M. L., 1998, Effect of ﬁrewood harvesting on
birds in a California oak-pine woodland, Journal of Wildlife Management
62(2):485-496

Amacher, G. S., Ersado, L., Hyde, W. F., Osorio, A., 2004, Tree planting in Tigray,
Ethiopia: the importance of human disease and water microdams, Agroforestry
Systems 60(3):211-225

An, L., He, G. M., Liang, Z., Liu, J. G., 2006, Impacts of demographic and
socioeconomic factors on spatio-temporal dynamics of panda habitat,
Biodiversity and Conservation 15(8):2343—2363

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
multidisciplinary and multiscale integration, Annals of the Association of
American Geographers 95(1):54-79

An, L., Liu, J. G., Ouyang, Z. Y., Linderman, M., Zhou, S. Q., Zhang, H. M., 2001,
Simulating demographic and socioeconomic processes on household level and
implications for giant panda habitats, Ecological Modelling 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, Ecological Economics 42(3):445-457

An, L., Mertig, A. G., Liu, J. G., 2003, Adolescents leaving parental home:
Psychosocial correlates and implications for conservation, Population and
Environment 24(5):415-444

141

Arambiza, E., Painter, M ., 2006, Biodiversity conservation and the quality of life of
indigenous people in the Bolivian Chaco, Human Organization 65(1):20-34

Archabald, K., Naughton-Treves, L., 2001, Tourism revenue-sharing around national
parks in Western Uganda: early efforts to identify and reward local communities,
Environmental Conservation 28(2): 135-149

Armstrong, E., Ball, J ., Bodoff, S., Carson, D. 8., Evans, 1., Green, D., Haase, K.,
Jendrock, E., 2005, The J2EE 1.4 Tutorial

Arnold, J. E. M., Kohlin, G., Persson, R., 2006, Woodfuels, livelihoods, and policy
interventions: Changing perspectives, World Development 34(3):596-611

Axinn, W. G., Pearce, L. D., Ghimire, D., 1999, Innovations in life history calendar
applications, Social Science Research 28(3):243-264

Batschelet, E., 1981, Circular statistics in biology, Academic Press Inc., London,
England, pp. xvi, 371

Bearer, S., Linderman, M., Huang, J. Y., An, L., He, G. M., Liu, J. 0., 2008, Effects of
fuelwood collection and timber harvesting on giant panda habitat use, Biological
Conservation 141(2):385-393

Bearer, S. L., 2005, The Effects of Forest Harvesting on Giant Panda Habitat Use in
Wolong Nature Reserve, China, in: Fisheries and Wildlife, Michigan State
University, East Lansing, MI, pp. pp 173

Bengtsson, J ., Persson, T., Lundkvist, H., 1997, Long-term effects of logging residue
addition and removal on macroarthropods and enchytraeids, Journal of Applied
Ecology 34(4):]014-1022

Bennett, M. T., 2008, China's sloping land conversion program: Institutional innovation
or business as usual? , Ecological Economics 65(4):699-711

Bhargava, H. K., Power, D. J ., 2001, Decision support systems and Web technologies: a
status report, in: Proceedings of the 2001 Americas Conference on Information
Systems, Boston, MA, August 3-5, 2001

142

 

 

Bhargava, H. K., Power, D. J ., Sun, D., 2007a, Progress in Web-based decision support
technologies, Decision Support Systems 43(4):1083-1095

Bhargava, H. K., Power, D. J ., Sun, D., 2007b, Special issue of Decision Support
Systems on web-based decision support, Decision Support Systems
43(4): 1081-1082

Bhatt, B. P., Negi, A. K., Todaria, N. P., 1994, Fuelwood Consumption Pattern at
Different Altitudes in Garhwal Himalaya, Energy 19(4):465-468

Bhatt, B. P., Sachan, M. S., 2004, Firewood consumption pattern of different tribal
communities in Northeast India, Energy Policy 32(1): 1-6

Bisong, F. E., Ambo, M. E., Ogar, I. I., 2007, Agroforestry systems for mitigating the
problems of fuelwood scarcity in the Montane/ Savanna ecosystems: a case study
of Obudu Cattle Ranch, South-Eastern Nigeria, Global Journal of
Environmental Sciences 6(1):53-64

Boot, R. G. A., Gullison, R. E., 1995, APPROACHES TO DEVELOPING
SUSTAINABLE EXTRACTION SYSTEMS FOR TROPICAL FOREST
PRODUCTS, Ecological Applications 5(4):896-903

Brown, K., 2003, Three challenges for a real people-centred conservation, Global
Ecology and Biogeography 12(2):89-92

Carlsson, C., Turban, E., 2002, DSS: directions for the next decade, Decision Support
Systems 33(2): 105-1 10

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 clinical

assessment method for collecting retrospective event-history data, International
Journal of Methods in Psychiatric Research 6(2): 101-1 14

Chakhar, S., Mousseau, V., 2008, GIS—based multicritieria spatial modeling generic
framework, International Journal of Geographical Information Science

22(11):]159-1196

Chan, K. M. A., Pringle, R. M., Ranganathan, J ., Boggs, C. L., Chan, Y. L., Ehrlich, P.
R., Haff, P. K., Heller, N. E., Al-Khafaji, K., Macmynowski, D. P., 2007, When

143

agendas collide: human welfare and biological conservation, Conservation
Biology 21(1):59-68

Chen, L., Heerink, N., van den Berg, M., 2006, Energy consumption in rural China: A
household model for three villages in J iangxi Province, Ecological Economics
58(2):407-420

China's Ministry of Forestry and WF, 1989, {%hhiﬁﬁiﬁﬁﬁilﬁﬂﬂﬂIﬁé
(Conservation and management plan for giant pandas and their habitat), Beijing,
China. [in Chinese]

China lntemet Network Information Center, 2008, The 21St Status Report of China
lntemet Network Development

China.com.cn, 2007, It [are $1 war/hi: 2531 4‘ antennae ﬁﬁhl‘éiﬁ (2531
Nature Reserves Established in China (before the end of 2007)),
http://www.china.com.cn/txt/2008-05/23/content_l 5419786.htm, Last Access:
09-13-2008

Chokkalingam, U., Zaizhi, Z., Chunfeng, W., Toma, T., 2006, Learning lessons from
China's forest rehabilitation efforts: national level review and special focus on
Guangdong Province, Center for International Forestry Research (CIFOR),
Jakarta, Indonesia, pp. ix + 159 pp.

Chomitz, K. M., Griffiths, C., 2001, An economic analysis and simulation of woodfuel
management in the Sahel, Environmental and Resource Economics
19(3):285-304

Cooke, P., Kohlin, G., Hyde, W. F., 2008, Fuelwood, forests and community
management - evidence from household studies, Environment and Development
Economics 13:103-135

Daily, G. C., 1997, Nature's services: Societal dependence on natural ecosystems, Island
Press, Washington, DC, USA

Daily, G. C ., Matson, P. A., 2008, Ecosystem services: From theory to implementation,
Proceedings of the National Academy of Sciences 105(28):9455-9456

144

de Anguita, P. M., Alonso, E., Martin, M. A., 2008, Environmental economic, political
and ethical integration in a common decision-making framework, Journal of
Environmental Management 88(1): 1 54-164

Dobson, A., 2005, Monitoring global rates of biodiversity change: challenges that arise
in meeting the Convention on Biological Diversity (CBD) 2010 goals,
Philosophical Transactions of the Royal Society B—Biological Sciences
360(1454):229-24l

Dong, J. W., Bai, S. H., Jia, H. M., 1999, Evaluation of thinning methods of Japanese
Larch plantations, Inner Mongolia Forestry Science & Technology
(3-4):102-103. [in Chinese]

Engel, B. A., Choi, J.-Y., Harbor, J ., Pandey, S., 2003, Web-based DSS for hydrologic
impact evaluation of small watershed land use changes, Computers and
Electronics in Agriculture 39(3):24l-249

Engel, S., Palmer, C., 2008, Payments for environmental services as an alternative to
logging under weak property rights: The case of Indonesia, Ecological
Economics 65(4):799-809

Fargione, J ., Hill, J ., Tilman, D., Polasky, 8., Hawthorne, P., 2008, Land clearing and
the biofuel carbon debt, Science 319(5867):1235-1238

Farsi, M., F ilippini, M., Pachauri, S., 2007, Fuel choices in urban Indian households,
Environment and Development Economics 12:757-774

Forest Products Laboratory, 2004, Fuel Value Calculator, US Forest Service

Franklin, S. E., Moskal, L. M., Lavigne, M. B., Pugh, K., 2000, Interpretation and
classiﬁcation of partially harvested forest stands in the Fundy Model Forest
using multi-temporal Landsat TM digital data, Canadian Journal of Remote
Sensing 26(4):318-333

Goldman, R. L., Tallis, H., Kareiva, P., Daily, G. C., 2008, Field evidence that
ecosystem service projects support biodiversity and diversify options,
Proceedings of the National Academy of Sciences of the United States of
America 105(27):9445-9448

145

Hall, T. E., F arrell, T. A., 2001, Fuelwood depletion at wilderness campsites: extent and
potential ecological signiﬁcance, Environmental Conservation 28(3):241-247

Hardin, G., 1968, The tragedy of the commons, Science 162(3859):1243-1248

He, G., Chen, X., Bearer, S., Colunga, M., Mertig, A., An, L., Zhou, S. Q., Linderman,
M., Ouyang, Z., Gage, S., Liu, J ., in review, Spatial and Temporal Patterns of
Fuelwood Collection in Wolong Nature Reserve: Implications for Panda
Conservation

He, G. M., Chen, X. D., Liu, W., Bearer, 8., Zhou, S. Q., Cheng, L. Y., Zhang, H. M.,
Ouyang, Z. Y., Liu, J. G., in press, Distribution of Economic Beneﬁts from
Ecotourism: A Case Study of Wolong Nature Reserve for Giant Pandas in China,
Environmental Management

He, G. R., 1991, An survey of diseases and pests of Japanese larch plantations in
Wolong Nature Reserve, Sichuan Forestry Science and Technology 12(4). [in
Chinese]

He, N., Liang, C., Yin, X., 1996, Sustainable community development in Wolong
Nature Reserve, Ecological Economy 1:15-23. [in Chinese]

Heltberg, R., 2004, Fuel switching: evidence from eight developing countries, Energy
Economics 26(5):869-887

Heltberg, R., 2005, Factors determining household fuel choice in Guatemala,
Environment and Development Economics 10:337-361

Hiemstra—van der Horst, G., Hovorka, A. J ., 2008, Reassessing the "energy ladder":
Household energy use in Maun, Botswana, Energy Policy 36(9):3333-3344

Kaimowitz, D., Sheil, D., 2007, Conserving What and for Whom? Why Conservation
Should Help Meet Basic Human Needs in the Tropics, Biotropica
39(5):567-574

Kang, S., J ia, L., Peng, Z., He, B., 2007, Progress on the species choice and silvicultural
practices of fuel energy forest, World Forestry Research 20(3):27-33. [in
Chinese]

146

Kearns, F. R., Kelly, M., Tuxen, K. A., 2003, Everything happens somewhere: using
webGIS as a tool for sustainable natural resource management, Frontiers in
Ecology and the Environment 1(10):541-548

Kochenderfer, J. N., Smith, H. C., Crews, J. T., 1995, Effects of Fertilization on the
Growth and Development of a Japanese Larch Plantation in West-Virginia,
USDA Forest Service Northeastern Forest Experiment Station Research Paper

(700):1-&

Kohlin, G., Parks, P. J ., 2001, Spatial variability and disincentives to harvest:
Deforestation and fuelwood collection in south Asia, Land Economics
77(2):206-218

Kumar, J., 2008, Biomass productivity Acacia auriculiformis as an important renewable
energy resource, Proc. Nat. Acad. Sci. India Sect. B-Biol. Sci. 78:77-84

Kumar, R., Shahabuddin, G., 2005, Effects of biomass extraction on vegetation
structure, diversity and composition of forests in Sariska Tiger Reserve, India,
Environmental Conservation 32(3):248-259

Leach, G., 1992, The Energy Transition, Energy Policy 20(2):l 16-123

Leach, G., Meams, R., 1988, Beyond the Fuelwood Crisis: People, Land and Trees in
Africa, Earthscan, London

Li, C. Y., Zhou, S. D., Xian, D. J ., Cheng, Z. Y., Tian, Z. X., 1992, Study of the
Management of Wolong Nature Reserve, in: The Animal and Plant Resources
and Protection of Wolong Nature Reserve (Wolong Nature Reserve, Sichuan
Normal College, eds.), Sichuan Publishing House of Science & Technology,
Chengdu, pp. 343-372. [in Chinese]

Li, D. X., 1982, A preliminary growth analysis of plantations of Japanese larch
introduced in mountainous areas in Sichuan Province, Sichuan Forestry Science
and Technology (3):36-41. [in Chinese]

Li, S. Y., Hou, Y. M., Yang, N. Y., Quan, Y. S., 2000, A prelimary study on thinning of
forest stands for Japanese larch plantations in northern subtropical and subalpine
regions, Bulletin of Forest Science and T echonologyz20-22. [in Chinese]

147

 

Li, Z. M., Tang, R. S., Xia, C. F ., Luo, H. L., Hao, Z., 2005, Towards green rural energy
in Yunnan, China, Renewable Energy 30(2):99-108

Linde-Rahr, M., 2003, Property rights and deforestation: The choice of fuelwood source
in rural Viet Nam, Land Economics 79(2):217-234

Linderman, M. A., An, L., Bearer, 8., He, G. M., Ouyang, Z. Y., Liu, J. G., 2005,
Modeling the spatio-temporal dynamics and interactions of households,
landscapes, and giant panda habitat, Ecological Modelling 183(1):47-65

Liu, J ., Li, S., Ouyang, Z., Tam, C., Chen, X., 2008a, Ecological and socioeconomic
effects of China's policies for ecosystem services, Proceedings of the National
Academy of Sciences 105(28):9477-9482

Liu, J ., Linderman, M., Ouyang, Z., An, L., Yang, J., Zhang, H., 2001a, Ecological
degradation in protected areas: the case of Wolong Nature Reserve for giant
pandas, Science 292:98-101

Liu, J ., Ouyang, Z., Tan, Y., Yang, J ., Zhang, H., 1999a, Changes in human population
structure: implications for biodiversity conservation, Population and
Environment 21(1):45-58

Liu, J. G., Daily, G. C., Ehrlich, P. R., Luck, G. W., 2003, Effects of household
dynamics on resource consumption and biodiversity, Nature 421(6922):530-533

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., 2007a, Complexity of coupled
human and natural systems, Science 317(5844):1513-1516

Liu, J. G., Dietz, T., Carpenter, S. R., Folke, C., Alberti, M., Redman, C. L., Schneider,
S. H., Ostrom, E., Moran, E., Pell, A. N., Lubchenco, J ., Taylor, W. W., Ouyang,
Z., Deadman, P., Kratz, T., Provencher, W., 2007b, Coupled human and natural
systems, Ambio 36(8):639-649

Liu, J. G., Li, S. X., Ouyang, Z. Y., Tam, C., Chen, X. D., 2008b, Ecological and
socioeconomic effects of China's policies for ecosystem services, Proceedings of
the National Academy of Sciences of the United States of America
105(28):9477-9482

148

Liu, J. G., Linderman, M., Ouyang, Z. Y., An, L., Yang, J., Zhang, H. M., 2001b,
Ecological degradation in protected areas: the case of Wolong Nature Reserve
for giant pandas, Science 292:98-101

Liu, J. G., Ouyang, Z. Y., Tan, Y. C., Yang, J ., Zhang, H. M., 1999b, Changes in human
population structure: implications for biodiversity conservation, Population and
Environment 21(1):45-58

Liu, J. G., Ouyang, Z. Y., Taylor, W. W., Groop, R., Tan, Y. C., Zhang, H. M., 1999c,
A framework for evaluating the effects of human factors on wildlife habitats: the
case of giant pandas, Conservation Biology 13(6):]360-1370

Loucks, C. J ., Lii, Z., Dinerstein, B, Wang, H., Olson, D. M., Zhu, Q, Wang, D., 2001,
Giant pandas in a changing landscape, Science 294:1465

Lowry, G., 1984, Commentary: Conservation Realism, Science, Technology, & Human
Values 9(4):54-57

Lu, X. L., 2004, A WEB-GIS based urgent medical rescue CSCW system for SARS
disease prevention, in: Grid and Cooperative Computing, Pt I, pp. 91-98

Lil, Y., Chen, L., Fu, B., Liu, S., 2003, A framework for evaluating the effectiveness of

protected areas: the case of Wolong Biosphere Reserve, Landscape and Urban
Planning 63(4):213-223

Ma, H., Han, G. T., Hui, S. Y., 1998, Development of a mathematical timber growth
model for artiﬁcial forest stands of Japanese Larix Kaempferi., Journal of
Shenyang University (Natural Sciences Edition) (4):]02-104. [in Chinese]

MacKinnon, J ., DeWulf, R., 1994, Designing protected areas for giant pandas in China,
in: Mapping the diversity of nature (R. I. Miller, ed.), Chapman & Hall, London,
England, pp. 142-172

Madeschin, F., 1999, Short rotation forestry in central and northern Europe -
introduction and conclusions, Forest Ecology and Management 121(1-2):1-7

Madubansi, M., Shackleton, C. M., 2007, Changes in fuelwood use and selection
following electriﬁcation in the Bushbuckn'dge lowveld, South Africa, Journal of
Environmental Management 83(4):4l 6-426

149

Mahiri, I., Howorth, C., 2001, Twenty years of resolving the irresolvable: Approaches
to the fuelwood problem in Kenya, Land Degradation & Development
12(3):205-215

Malczewski, J ., 1999, GIS and Multicriteria Decision Analysis, J. Wiley & Sons, New
York, pp. 392

Malczewski, J ., 2006, GIS-based multicriteria decision analysis: a survey of the
literature, International Journal of Geographical Information Science
20(7):703—726

Mandondo, A., 2001, Use of woodland resources within and across villages in a
Zimbabwean communal area, Agriculture and Human Values 18(2): 1 77-1 94

Masera, O. R., Saatkamp, B. D., Kammen, D. M., 2000, From linear fuel switching to
multiple cooking strategies: A critique and alternative to the energy ladder
model, World Development 28(12):2083-2103

Mathiyalagan, V., Grunwald, S., Reddy, K. R., Bloom, S. A., 2005, A WebGIS and
geodatabase for Florida's wetlands, Computers and Electronics in Agriculture
47(1):69-75

Millennium Ecosystem Assessment, 2005, ECOSYSTEMS AND HUMAN
WELL-BEING: CURRENT STATE AND TRENDS, Island Press, Washington,
DC, pp. 815

Moy, O. L., 2005, The "Grain-To-Green" Program and Its Applications for
Conservation of Giant Panda Habitat in Wolong Nature Reserve, China, in:
Department of Ecology and Evolutionary Biology, Princeton University,
Princeton, New Jersey, pp. 59

Mysiak, J ., Giupponi, C., Rosato, P., 2005, Towards the development of a decision
support system for water resource management, Environmental Modelling &
Software 20(2):203-214

Neto, F., 2003, A new approach to sustainable tourism development: Moving beyond
environmental protection, Natural Resources Forum 27(3):212-222

150

N gugi, A. W., 1988, Cultural-Aspects of Fuelwood Shortage in the Kenyan Highlands,
Journal of Biogeography 15(1): 1 65-1 70

Nygard, R., Elfving, B., 2000, Stem basic density and bark proportion of 45 woody
species in young savanna coppice forests in Burkina Faso, Annals of Forest
Science 57(2): 143-1 53

Nygard, R., Sawadogo, L., Elfving, B., 2004, Wood-fuel yields in short-rotation
coppice growth in the north Sudan savanna in Burkina F aso, Forest Ecology and
Management 189(1-3):77-85

Orr, D. W., 2003, Walking North on a Southbound Train, Conservation Biology
l7(2):348-351

Paassen, A. v., Roetter, R. P., Keulen, H. v., Hoanh, C. T., 2007, Can computer models
stimulate learning about sustainable land use? Experience with LUPAS in the
humid (sub-)tropics of Asia, Agricultural Systems 94(3):874-887

Palmer, G., 2000, Java Programmer's Reference, Wrox Press, Chicago, IL, pp. 1227

Pattanayak, S. K., Sills, E. O., Kramer, R. A., 2004, Seeing the forest for the fuel,
Environment and Development Economics 9: 155-1 79

Phua, M. H., Minowa, M., 2005, A GIS-based multi-criteria decision making approach

to forest conservation planning at a landscape scale: a case study in the Kinabalu
Area, Sabah, Malaysia, Landscape and Urban Planning 71(2-4):207-222

Pinzon, Z. S., Ewel, K. C., Putz, F. E., 2003, Gap formation and forest regeneration in a
Micronesian mangrove forest, Journal of Tropical Ecology 19:143-153

Poch, M., Comas, J ., Rodriguez-Roda, 1., sanchez-Marre, M., Cortes, U., 2004,
Designing and building real environmental decision support systems,
Environmental Modelling & Software l9(9):857-873

Population Action International, 1999, One in Three People Lives in Forest-scarce
Countries, Population Action International, Washington DC

151

Prasad, R., Maithel, S., Mirza, A., 2001, Renewable energy technologies for fuelwood
conservation in the Indian Himalayan region, Sustain. Dev. 9(2):]03-108

Rauscher, H. M., Reynolds, K., Vacik, H., 2005, Decision-support systems for forest
management, Computers and Electronics in Agriculture 49(1):1-5

Ravallion, M., Chen, S. H., 2005, Hidden impact? Household saving in response to a
poor-area development project, J. Public Econ. 89(1 1-12):2183-2204

Redford, K., Sanjayan, M. A., 2003, Retiring Cassandra, Conservation Biology
l7(6):1473-l474

Reid, D. G., Taylor, A. H., Hu, J. C., Qin, Z. S., 1991, Environmental-Inﬂuences on
Bamboo Bashania-Fangiana Growth and Implications for Giant Panda
Conservation, Journal of Applied Ecology 28(3):855-868

Richardson, J ., Bjorheden, R., Hakkila, P., Lowe, A. T., Smith (eds), C. T., 2002,
Bioenergy from Sustainable Forestry: Guiding Principles and Practice, Kluwer
Academic Publishers

Rijal, K., 1999, Renewable energy policy options for mountain communities:
Experiences from China, India, Nepal and Pakistan, Renewable Energy
l6(1-4):ll38-1142

Robinson, J. G., 2006, Conservation biology and real-world conservation, Conservation
Biology 20(3):658-669

Ronnback, P., Crona, B., Ingwall, L., 2007, The return of ecosystem goods and services
in replanted mangrove forests: perspectives from local communities in Kenya,
Environmental Conservation 34(4):313-324

Rosenstock, S. S., 1998, Inﬂuence of Gambel oak on breeding birds in ponderosa pine
forests of Northern Arizona, Condor 100(3):485-492

Rossi, P., Pecci, A., Amadio, V., Rossi, 0., Soliani, L., 2008, Coupling indicators of
ecological value and ecological sensitivity with indicators of demographic
pressure in the demarcation of new areas to be protected: The case of the
Oltrepo Pavese and the Ligurian-Emilian Apennine area (Italy), Landscape and

Urban Planning 85(1):]2-26
1 52

Sagar, R., Singh, J. S., 2004, Local plant species depletion in a tropical dry deciduous
forest of northern India, Environmental Conservation 31(1):55-62

Salafsky, N., Cauley, H., Balachander, G., Cordes, B., Parks, J ., Margoluis, C., Bhatt, S.,
Encamacion, C., Russell, D., Margoluis, R., 2001, A systematic test of an
enterprise strategy for community-based biodiversity conservation,

Conservation Biology 15(6): 1585-1 595

Samant, S. S., Dhar, U., Rawal, R. S., 2000, Assessment of fuel resource diversity and
utilization patterns in Askot Wildlife Sanctuary in Kumaun Himalaya, India, for
conservation and management, Environmental Conservation 27(1):5-l3

Sante, l., Crecente, R., Miranda, D., Tourino, J ., Canzobre, F., Doallo, R., 2004, A GIS
web-based tool for the management of the PGI potato of Galicia, Computers and
Electronics in Agriculture 44(2): 161-1 71

Sayer, J., Chokkalingam, U., Poulsen, J ., 2004, The restoration of forest biodiversity
and ecological values, Forest Ecology and Management 201(1):3-1 l

Schaller, G. B., Hu, J ., Pan, W., Zhu, J ., 1985, The Giant Pandas of Wolong, University
of Chicago Press, Chicago, Illinois, USA

Schwankert, S., 2008, China's lntemet population tops 200 million, InfoWorld

Sedjo, R. A., 2008, Biofuels: Think Outside the Cornﬁeld, Science
320(5882):l420-1421

Shanavas, A., Kumar, B. M., 2003, Fuelwood characteristics of tree species in
homegardens of Kerala, India, Agroforestry Systems 58(1):] 1-24

Sharma, U. R., 1990, An overview of park-people interactions in Royal Chitwan
National Park, Nepal, Landscape and Urban Planning 19(2): 133-144

Shen, Y. Q., Liao, X. C., Yin, R. S., 2006, Measuring the socioeconomic impacts of
China's Natural Forest Protection Program, Environment and Development
Economics 11:769-788

153

 

Shunula, J. P., 2002, Public awareness, key to mangrove management and conservation:
the case of Zanzibar, T recs-Structure and Function l6(2-3):209-212

Sichuan Forestry Survey Institute and Wolong Nature Reserve, 2000a, Plan for Natural
Forest Conservation Program in Wolong Nature Reserve (unpublished, in
Chinese)

Sichuan Forestry Survey Institute and Wolong Nature Reserve, 2000b, [El l l l E E1717 5
%lﬁ‘ci’f‘EiﬁﬁiﬁﬁiﬁbeEﬁﬁ’ghﬂﬂi‘lﬁﬁ (Plan for Natural Forest

Conservation Program in Wolong Nature Reserve). [in Chinese]

Solh, M., Amri, A., Ngaido, T., Valkoun, J ., 2003, Policy and education reform needs
for conservation of dryland biodiversity, Journal of Arid Environments
54(1):5-1 3

State Forestry Administration, 2005, The 3rd National Survey Report on Giant Panda in
China, China Science Press, Beijing, pp. 355

Su, Y. M., Liu, X. L., He, F., Liu, H. Y., Wang, Q. S., 2003, Study on nutrient elements
of Larix Kaempferi plantations and the distributions, Journal of Sichuan
Forestry Science & Technology 24(3):35-39. [in Chinese]

Sun Microsystems Inc., 2002, Model-View-Controller,
http://j ava.sun.com/blueprints/pattems/MVC .html, Last Access: 07-12-2006

Tabuti, J. R. S., Dhillion, S. S., Lye, K. A., 2003, Firewood use in Bulamogi County,
Uganda: species selection, harvesting and consumption patterns, Biomass &
Bioenergy 25(6):581-596

Tang, Z., Engel, B. A., Choi, J ., Sullivan, K., Sharif, M., Lim, K. J ., 2004, A Web-based
DSS for erosion control structure planning, Applied Engineering in Agriculture
20(5):707-714

Taylor, A. H., Qin, Z. S., 1993, Bamboo Regeneration after Flowering in the Wolong
Giant Panda Reserve, China, Biological Conservation 63(3):231-234

The Editorial Board of Forests in Sichuan, 1992, Forests in Sichuan, Forestry
Publishing House, Beijing, China. [in Chinese]

154

Top, N., Mizoue, N., Kai, S., Nakao, T., 2004, Variation in woodfuel consumption
patterns in response to forest availability in Kampong Thom Province,
Cambodia, Biomass & Bioenergy 27(1):57-68

Trac, C. J ., Harrell, S., Hinckley, T. M., Henck, A. C., 2007, Reforestation programs in
southwest China: Reported success, observed failure, and the reasons why,
Journal of Mountain Science 4(4):275-292

Turner, W. R., Brandon, K., Brooks, T. M., Costanza, R., da Fonseca, G. A. B., Portela,
R., 2007, Global conservation of biodiversity and ecosystem services,
BioScience 57(1 0):868-8 73

Uchida, E., Xu, J. T., Rozelle, S., 2005, Grain for green: Cost-effectiveness and
sustainability of China's conservation set-aside program, Land Economics
81(2):247-264

UNESCO World Heritage Centre, 2006, Sichuan Giant Panda Sanctuaries

Van Rijsoort, J., Zhang, J. F., 2005, Participatory resource monitoring as a means for
promoting social change in Yunnan, China, Biodiversity and Conservation
14(1 1):2543-2573

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, Ecological
Applications 17(4): 1019-1030

Vitousek, P. M., Mooney, H. A., Lubchenco, J ., Melillo, J. M., 1997, Human
domination of Earth's ecosystems, Science 277(5325):494-499

Vodak, M. C., Cupp, D. C., Hokans, R. H., 1986, Direct Net Energy Values from
Energy Inputs and Outputs for Fuelwood Harvest in Appalachian Hardwoods,
Journal of Environmental Management 22(4):319-330

Wang, D., Li, 8., Sun, S., Wang, H., Chen, A., Li, S., Li, J., Lu, Z., 2008, Turning
Earthquake Disaster into Long-Term Beneﬁts for the Panda, Conservation
Biology 22(5):]356-1360

155

 

Wang, G. Y., Innes, J. L., Lei, J. F., Dai, S. Y., Wu, S. W., 2007, Ecology - China's
forestry reforms, Science 318: 1556-1 557

Wang, X. W., Tian, Z. H., Yan, W. B., Luo, W. J., Meng, X. L., 1996, Study on Yield
Table of Artiﬁcial Forest Stands of Japanese Larch, Journal of Liaoning
Forestry Science & Technology (l):32-36. [in Chinese]

Weyerhaeuser, H., Wilkes, A., Kahrl, F., 2005, Local impacts and responses to regional
forest conservation and rehabilitation programs in China's northwest Yunnan
province, Agricultural Systems 85(3):234-253

Wolong Administration Bureau Department of Social and Economic Development,
2006, Annual report on the rural economy of Wolong Nature Reserve, Wolong
Administration Bureau. [in Chinese]

Wolong Nature Reserve, 1998a, Master Plan for Wolong Nature Reserve (in Chinese)

Wolong Nature Reserve, 1998b, [EI‘IZ El %1'1 1}" [X E} 1715 I'll 3211 (Master Plan for
Wolong Nature Reserve). [in Chinese]

Wolong Nature Reserve, 2000a, ENE“. 2000 {15A C1 1"?" fl (Wolong population census
in 2000). [in Chinese]

Wolong Nature Reserve, 2000b, Efﬁiﬂiﬁ—‘d 122 BE ‘lL @111 (Compilation of codes and
regulations in Wolong). [in Chinese]

Wolong Nature Reserve, 2005a, Development of Wolong Nature Reserve (in Chinese),
Sichuan Science and Technology Press, Chengdu, pp. 154

Wolong Nature Reserve, 2005b, ﬁl‘jﬁﬁlﬁ 93 (History of the Development of
Wolong Nature Reserve), Sichuan Science Publisher, Chengdu, China, pp. 154.
[in Chinese]

Wong, 1. W., Bloom, R., McNicol, D. K., Fong, P., Russell, R., Chen, X., 2007, Species
at risk: Data and knowledge management within the WILDSPACE(TM)
Decision Support System, Environmental Modelling & Software 22(4):423-430

156

 

Wunder, S., Engel, S., Pagiola, S., 2008, Taking stock: A comparative analysis of
payments for environmental services programs in developed and developing
countries, Ecological Economics 65(4):834-852

Xu, J. T., Yin, R. S., Li, Z., Liu, C., 2006a, China's ecological rehabilitation:
Unprecedented efforts, dramatic impacts, and requisite policies, Ecological
Economics 57(4):595-607

Xu, Z. G., Xu, J. T., Deng, X. Z., Huang, J. K., Uchida, E., Rozelle, S., 2006b, Grain for
green versus grain: Conﬂict between food security and conservation set-aside in
China, World Development 34(1): 1 30-148

Yan, W. X., Liang, H. 0,19%, A preliminary study of the introduction of Japanese
larch into Sichuan, Sichuan Forestry Science and Technology 14(3):8-14. [in
Chinese]

Yang, Y. B., 1989, Ecological Study of Sichuan Forest, Chengdu Science &
Technology Press, Chengdu. [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, Mountain
Research and Development 23(4):345-3 52

Zhang, H., Li, 1)., Wei, R. 1)., Tang, C., Tu, J., 1997, atnxntamaahaaraua
iZUEE (Advances in conservation and studies on reproductivity of giant pandas
in Wolong), Sichuan Journal of Zoology (16):31-33

Zhang, L., Huang, Y., Luo, T. X., Dai, Q., Deng, K. M., 2005, Age effects on stand
biomass allocations to different components: a case study in forests of

Cunninghamia Lanceolata and Pinus Massoniana, Journal of the Graduate
School of the Chinese Academy of Sciences 22(2): 1 70-178. [in Chinese]

Zhang, L., Tu, Q., Mol, A. P. J., 2008, Payment for environmental services: The sloping
land conversion program in Ningxia autonomous region of China, China World
Econ. 16(2):66-81

Zhang, R, Shao, G., Zhao, G., Master, D. C. L., Parker, G. R., Dunning Jr., J. B., Li, Q.,
2000, China's Forest Policy for the 2lst Century, Science 288:2135-2136

157

Zhu, C., Feng, G., 2003, at: narranmw ﬁrsﬁﬁaelﬁl 1713‘: (A Case Study on
China's Policy of Converting Steep Cultivated Land to Forest or Grassland),
China's Science Press, Beijing, China. [in Chinese]

158