Scholarly article on topic 'Assessing the likelihood of widespread landholder adoption of afforestation and reforestation projects'

Assessing the likelihood of widespread landholder adoption of afforestation and reforestation projects Academic research paper on "Agriculture, forestry, and fisheries"

CC BY-NC-ND
0
0
Share paper
Academic journal
Global Environmental Change
OECD Field of science
Keywords
{Adoption / Afforestation / Reforestation / "Carbon sequestration" / Mitigation}

Abstract of research paper on Agriculture, forestry, and fisheries, author of scientific article — Jacki Schirmer, Lyndall Bull

Abstract Tree planting for carbon sequestration is a commonly proposed climate change mitigation strategy, with afforestation projects forming part of voluntary and mandatory carbon offset trading schemes. Afforestation is often promoted as a new economic opportunity for private landholders. While multiple studies have identified physical and economic opportunities for afforestation, few have examined the willingness of private landholders to adopt afforestation, and the factors that influence this willingness. We examine this using data from a survey of Australian landholders. The willingness of landholders to adopt afforestation for carbon sequestration varies substantially depending on how this afforestation is designed and implemented: landholders prefer small plantings on less productive land, which minimise the disturbance afforestation presents to land management, and to landholder values about appropriate uses of agricultural land. Landholders are less willing to consider afforestation if it involves planting the large areas required by many current carbon afforestation schemes. Willingness to adopt afforestation is influenced in particular by landholder's perceptions of its potential to provide a diversified income stream, and its impacts on flexibility of land management. More broadly, it is influenced by their views about the social acceptability of afforestation, particularly whether the landholder believes trees should be planted on agricultural land, and how they believe others in the community view afforestation. Our results suggest that widespread adoption requires designing afforestation so it (i) provides a range of socio-economic benefits that go beyond provision of income; (ii) minimises disruption to land management flexibility; and (iii) is compatible with landholder beliefs about appropriate use of agricultural land.

Academic research paper on topic "Assessing the likelihood of widespread landholder adoption of afforestation and reforestation projects"

Contents lists available at ScienceDirect

Global Environmental Change

journal homepage www.elsevier.com/locate/gloenvcha

Global Environmental Change

Assessing the likelihood of widespread landholder adoption of afforestation and reforestation projects^

Jacki Schirmera'b,c'*, Lyndall Bulla'b

a Fenner School of Environment and Society, Building 48, Australian National University, ACT 0200, Australia b Cooperative Research Centre for Forestry, Private Bag 12, Hobart, TAS 7001, Australia c Centre for Research and Action in Public Health, Building 22, University of Canberra, 0200, Australia

CrossMark

ARTICLE INFO

ABSTRACT

Article history:

Received 28 November 2011

Received in revised form 2 November 2013

Accepted 6 November 2013

Keywords:

Adoption

Afforestation

Reforestation

Carbon sequestration

Mitigation

Tree planting for carbon sequestration is a commonly proposed climate change mitigation strategy, with afforestation projects forming part of voluntary and mandatory carbon offset trading schemes. Afforestation is often promoted as a new economic opportunity for private landholders. While multiple studies have identified physical and economic opportunities for afforestation, few have examined the willingness of private landholders to adopt afforestation, and the factors that influence this willingness. We examine this using data from a survey of Australian landholders. The willingness of landholders to adopt afforestation for carbon sequestration varies substantially depending on how this afforestation is designed and implemented: landholders prefer small plantings on less productive land, which minimise the disturbance afforestation presents to land management, and to landholder values about appropriate uses of agricultural land. Landholders are less willing to consider afforestation if it involves planting the large areas required by many current carbon afforestation schemes. Willingness to adopt afforestation is influenced in particular by landholder's perceptions of its potential to provide a diversified income stream, and its impacts on flexibility of land management. More broadly, it is influenced by their views about the social acceptability of afforestation, particularly whether the landholder believes trees should be planted on agricultural land, and how they believe others in the community view afforestation. Our results suggest that widespread adoption requires designing afforestation so it (i) provides a range of socio-economic benefits that go beyond provision of income; (ii) minimises disruption to land management flexibility; and (iii) is compatible with landholder beliefs about appropriate use of agricultural land.

© 2013 The Authors. Published by Elsevier Ltd. All rights reserved.

1. Introduction

The planting of trees to sequester carbon is a commonly advocated climate change mitigation strategy. It is included in the Clean Development Mechanism (CDM) of the United Nations Framework Convention on Climate Change (UNFCC), and afforestation and reforestation projects form part of several voluntary and mandatory carbon offset trading schemes worldwide (Diaz et al., 2011). Afforestation refers to planting trees on land not afforested in recent history (usually at least 50 years), while reforestation

§ This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial-No Derivative Works License, which permits non-commercial use, distribution, and reproduction in any medium, provided the original author and source are credited.

* Corresponding author at: Centre for Research and Action in Public Health, Building 22, University of Canberra, 0200, Australia. Tel.: +61 2 6201 2785/428 254 948; fax: +61 2 6125 0746.

E-mail addresses: Jacki.schirmer@canberra.edu.au (J. Schirmer), Lyndall.bull@anu.edu.au (L. Bull).

refers to the replanting of trees on more recently deforested land (Hamilton et al., 2010). Land must have been cleared prior to 1989 to be eligible under the CDM (Hamilton et al., 2010).

Despite widespread promotion in recent years, afforestation and reforestation projects have not gained the traction hoped for by their advocates. By September 2011 they made up only 0.75% of registered CDM projects, with approximately 403,000 hectares of trees planted by 241 afforestation and reforestation projects servicing mandatory and voluntary markets worldwide (Diaz et al., 2011). Even assuming this is an underestimate, the figure is low compared to the 264 million hectares (MHa) of planted forests globally (FAO, 2010), and the varied estimates of a need for plantings in the order of tens to hundreds MHa if afforestation and reforestation are to make a meaningful contribution to addressing human-induced climate change (e.g. Cannell, 2003; Zomer et al., 2008). The relatively small contribution of afforestation and reforestation to carbon markets may result from many factors, including market barriers, high upfront costs, and the long time before trees sequester large amounts of carbon (Thomas et al., 2010; Diaz et al., 2011). However, these projects are growing in their contribution to traded carbon as

0959-3780/$ - see front matter © 2013 The Authors. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.gloenvcha.2013.11.009

early plantings reach an age at which they sequester significant levels of carbon: afforestation- and reforestation-grown carbon constituted 10% of transactions in voluntary carbon markets in 2011 (Peters-Stanley and Hamilton, 2012).

From this point on, we use the term 'afforestation' in this paper to refer to afforestation and reforestation for carbon sequestration; the content of the paper should be assumed to apply to both unless otherwise specified. When referring to the planting of trees for purposes other than carbon sequestration, we do not use the term 'afforestation', to avoid confusion.

Afforestation is an often controversial aspect component of the CDM. Critics argue that it may divert attention from the need to reduce greenhouse gas emissions and consumption, exploit developing countries, have negative environmental impacts, or be ineffective at mitigating carbon (Ciscell, 2010). Proponents argue that afforestation has the potential not only to achieve significant mitigation, but to provide livelihood benefits for landholders in both developing and developed countries. Irrespective of perceptions of costs and benefits, the reality is that afforestation is a part of climate mitigation policy, and likely to remain so into the future (Corbera and Brown, 2010). Given this, it is important to consider how to ensure afforestation has benefits for landholders, both to ensure that they are not adversely affected by afforestation, and to increase the likelihood that they will adopt afforestation and thus increase its contribution to climate mitigation.

To be successful as a mitigation strategy, afforestation projects must achieve net sequestration after taking into account issues such as carbon outcomes under previous land uses and leakage. Considerable attention has been given to these issues (e.g. Murray et al., 2004; Sathaye and Andrasko, 2007; van Minnen et al., 2008), as well as to identifying the total area of land physically suitable for afforestation projects (e.g. Zomer et al., 2008), and the area of land likely to achieve a positive economic return for landholders under different carbon prices (Benitez et al., 2007; Winsten et al., 2011; Palmer and Silber, 2012; Yemshanov et al., 2012; Paul et al., 2013; Polglase et al., 2013). Several studies have examined whether afforestation is a cost effective mitigation strategy compared to other options climate mitigation options (Richards and Stokes, 2004; van Kooten et al., 2004), particularly compared to use of bioenergy (Rootzen et al., 2010; Kallio et al., in press).

Less attention has been given to the social and economic challenges of achieving large scale adoption of afforestation. A small number of studies have identified social, economic, and institutional factors that may limit the success of afforestation projects, focusing on the potential social, economic and environmental impacts of afforestation, transaction costs, land tenure issues, and competition for use of land (Nilsson and Schopfhauser, 1995; Cannell, 2003; Jindal et al., 2008; Thomas et al., 2010). These have built some understanding of the broader economic and social ramifications of afforestation, but Jindal et al. (2012) identified that there remains a need to examine the benefits and costs of afforestation beyond simple provision of economic return to landholders.

We argue that there is an additional need: to better understand what the various social and economic costs and benefits of afforestation mean for the likelihood of widespread adoption of afforestation by landholders (Bozmoski and Hultman, 2010). Theoretical estimates of afforestation potential, based on biophysical and economic viability, have limited usefulness unless accompanied by an understanding of whether and when landholders are willing to consider afforestation. In many countries, afforestation can only occur on a large scale via tree planting on privately owned land. This means afforestation will only be successful if private landholders are convinced to participate in afforestation projects, and to maintain their participation over long periods. The factors influencing landholder adoption of afforestation are not well understood, representing a significant gap in

current knowledge regarding afforestation's likely success as a mitigation strategy (Bull and Thompson, 2011). The few studies undertaken have typically assumed adoption will depend on the economic returns landholders achieve from afforestation, or on sociodemographic characteristics of landholders (van Kooten et al., 2002; Shaikh et al.,2007; Cacho et al., 2008; Torres et al., 2010), but have not formally tested these assumptions.

Meanwhile, an extensive literature on landholder adoption of new technologies and conservation practices has studied the agricultural landscapes where afforestation projects need to gain traction. This literature shows that landholder decisions are not driven solely by economic considerations, but rather depend on a wide range of factors, including the relative advantage of the new activity (e.g. its perceived costs and benefits), individual and social learning processes, extension efforts, sociodemographic and farm characteristics, influence of social networks, the supportiveness of institutional frameworks, and access to the physical, natural and financial resources needed for adoption, amongst other factors (Pannell et al., 2006; Prokopyetal.,2008; Ommani etal.,2009; Bull and Thompson, 2011). This suggests a need to examine the likely success of afforestation from an adoption perspective. Questions such as the following need to be asked to complement existing assessments of biophysical and economic feasibility of afforestation: how likely is it that landholders will be willing to adopt afforestation? Under what circumstances? What factors are most important in affecting willingness to adopt? How can afforestation be designed to achieve the goal of sequestering carbon while also being attractive to landholders?

In this paper, we examine the likelihood of widespread adoption of afforestation, and the factors that influence the willingness of landholders to adopt, using the results of a survey of landholders in Australia. First, we review existing studies on afforestation, and on tree planting and landholder adoption of new practices on agricultural land more broadly, to identify factors likely to influence adoption. We then describe the methods used to survey landholders. Our results analyse the likelihood of widespread adoption of afforestation, and how landholder's willingness to adopt varies depending on the design of afforestation. We then identify the factors that have the greatest influence on willingness to adopt. Our discussion focuses on the implications of our findings for those seeking to encourage adoption of afforestation on private land. By examining the role of landholder perceptions and beliefs, we contribute a new perspective that is critical to understanding whether afforestation is likely to be adopted at scales that enable it to make a meaningful contribution to offsetting greenhouse gas emissions.

2. Factors likely to influence landholder adoption of afforestation

An extensive literature has examined when and why landholders are willing to adopt new land management practices in general, and can be used to inform studies of willingness to adopt afforestation. Rather than reviewing the highly diverse literature on landholder adoption of new practices in its entirety, we identified factors likely to be relevant to landholder adoption of afforestation by drawing on three types of literature.

First, we examined recent meta-studies of factors influencing landholder adoption of new land management activities, focused on Australian (Pannell et al., 2006), American (Prokopy et al., 2008; Baumgart-Getz et al., 2012), and international studies (Knowler and Bradshaw, 2007). Second, we reviewed previous studies that have proposed factors likely to influence landholder adoption of afforestation (e.g. Bull and Thompson, 2011), or reviewed socioeconomic costs and benefits to landholders of afforestation (e.g. Jindal et al., 2012). Third, we reviewed studies that have examined

landholder adoption of (i) tree planting for commercial wood production (e.g. Race and Curtis, 2007; McGinty et al., 2008) and (ii) landholder adoption of on-farm tree planting for environmental purposes (e.g. Schirmer et al., 2012). Previous research on adoption of tree planting for commercial wood production has relevant lessons as this type of tree planting often takes place on the same cleared agricultural land sought by carbon sequestration projects; utilises the same types of mechanisms (e.g. sharefarming, lease or land purchase); and to be successful requires landholders to engage with both tree planting activity and a commercial market. We included only a small number of studies examining tree planting for environmental purposes, as this type of planting differs from afforestation plantings in many respects: there is no need to interface with a market; there are no required timeframes; and the landholder has more latitude in selecting site, species, and methods than for either afforestation projects or tree planting for wood production.

Based on our review, we identified the following factors as being highly likely to influence landholder willingness to adopt afforestation: design of afforestation; social acceptability of afforestation; socio-demographic attributes of landholders and their properties; landholder skills, knowledge and experience relevant to afforestation; and perceptions of attributes of afforestation. Each is described below.

2.1. Design of afforestation

The way afforestation is designed - for example, the tree species planted, area and lay out of planting, and other characteristics -will influence adoption. In particular, the scale of planting, type of land planted, and presence of co-benefits appear important. Studies examining social impacts of tree planting for commercial wood production have established a preference for smaller scale plantings on marginal land, and for planting to be undertaken by farmers rather than by afforestation companies often viewed as 'outsiders' (Barlow and Cocklin, 2003; Schirmer, 2007; Williams, 2011). Small-scale afforestation by landholders is rarely associated with social conflict in rural communities, whereas large-scale afforestation of farm land involving purchase of land by organisations other than farmers is often associated with social conflict and dissent in rural communities (Schirmer, 2007). The issue of encouraging small- versus large-scale trees planting has received considerable attention in the afforestation literature, with concerns raised about the environmental and social impacts of larger scale afforestation plantings (Smith and Scherr, 2003; Razak et al., 2009; Palmer and Silber, 2012). Despite these concerns, the majority of methodologies approved under the CDM are for large-scale afforestation projects (Razak et al., 2009). Other evidence suggests that landholders may prefer trees to be established on less productive (marginal) land, which has fewer opportunity costs than more productive land, an issue identified as relevant in regions including Australia, Africa and Europe (Schirmer, 2007; McDonagh et al., 2010; Reynolds, 2012); and prefer trees to have co-benefits beyond carbon payments, such as provision of commercial return from wood production (Anderson and Zerriffi, 2012; Palmer and Silber, 2012).

2.2. Social acceptability of afforestation

Landholders will be more willing to adopt new technologies if they view them as socially acceptable, meaning they are compatible with a landholder's beliefs and values, and with social norms about appropriate rural land management practices (Pannell et al., 2006; Prokopy et al., 2008; Bull and Thompson, 2011). The findings of some afforestation studies, despite not directly examining adoption, provide evidence that supports this

hypothesis: Bozmoski and Hultman (2010) found that adoption of afforestation in Tanzania was facilitated by a landholder's social networks, suggesting that if that network is unsupportive of afforestation, adoption is less likely. Boyd et al. (2007) identified that implementation of afforestation projects in Brazil and Bolivia was limited by low acceptance of afforestation, and social conflict about it.

Given the importance of social acceptance, the critical question is which social values most determine social acceptance. We identified four social values as having theoretical relevance. First, previous studies have identified that a landholders' beliefs about whether human activities are changing the climate, and whether trees are an effective method of sequestering carbon, are potentially influential (Blennow and Persson, 2009). Second, beliefs about whether good agricultural land should be used to grow trees have been identified as important in several countries in previous studies of tree planting for commercial wood production (Schirmer, 2007; McDonagh et al., 2010). Third, landholder views about the responsibility of farmers to manage land for long-term sustainability and the 'greater good' may be influential on adoption decisions (Schirmer et al., 2012). Fourth, the views of others in the landholder's social networks and broader community about the acceptability of afforestation are likely to be relevant (Pannell et al., 2006).

The issues of (i) 'good agricultural land' and (ii) landholder's obligations to manage land sustainably were considered relevant, as a large literature has highlighted these two issues as central to landholder's values in many countries. Productivist values, which imbue growing food on 'good agricultural land' as an activity with a moral worth that goes beyond economic return, often dominate the social norms of landholders (Burton and Wilson, 2006). This results in landholders elevating food production as a priority over other forms of land use. Burton and Wilson (2006) argue that productivist values tend to dominate stewardship values, in which landholders view themselves as having an obligation to manage land in an environmentally sustainable manner (Flick et al., 2010). The primacy of food production to landholder values has been found to act as a barrier to adoption of tree planting in a number of countries (McDonagh et al., 2010), and concerns about afforestation displacing food production are well established (e.g. Smith and Scherr, 2003). Adoption of afforestation is thus likely to depend in part on whether farmers view it as compatible with their values and beliefs around the primacy of food production, and their stewardship obligations.

The views of others were also considered relevant. The role of social capital in adoption of new practices by landholders is increasingly recognised, with adoption often strongly linked to the acceptability of a practice within the social networks a landholder is involved in (Knowler and Bradshaw, 2007), an issue identified as affecting adoption of tree planting for wood production (Race and Curtis, 2007). Commercial tree planting for wood production is commonly associated with contention in rural communities; conflict over the social, economic and environmental impacts of this type of tree planting has been recorded in more than 40 countries worldwide (Cossalter and Pye-Smith, 2003; Schirmer, 2007; Gerber, 2011). Similar concerns have been raised about afforestation in some locations (e.g. Boyd et al., 2007). The emergence of social conflict is suggestive of low social acceptance of afforestation, and may be linked to lower willingness to adopt by landholders, due to concern that they will be viewed negatively in their community if they take on a socially unacceptable enterprise on their land.

2.3. Socio-demographic attributes of landholders and their properties

Many adoption studies examine the socio-demographic characteristics of landholders (e.g. age, education level, income, and

goals and motivations) and of their property (e.g. property size, length of ownership, enterprise type), to see if these characteristics influence willingness to adopt particular practices. While some meta-analyses of adoption studies suggest income is significantly related to adoption in many studies (Baumgart-Getz et al., 2012), including willingness to adopt tree planting for commercial wood production (McGinty et al., 2008), other meta-analyses have found little or no consistent relationship between socio-demographic characteristics of landholders and adoption behaviour (Knowler and Bradshaw, 2007). Similarly, the characteristics of the land managed by landholders is sometimes, but not always, related to willingness to adopt, particularly farm size and tenure (Knowler and Bradshaw, 2007; Baumgart-Getz et al., 2012).

2.4. Landholder skills, knowledge and experience relevant to afforestation

Adoption of new practices is influenced by the landholder's confidence in their existing skills, knowledge and experience relevant to the new practice being considered, and their ability to access desired skills and knowledge (Pannell et al., 2006). Being able to access quality information about new practices has been consistently identified as a significant influence on adoption behaviour in multiple studies (Baumgart-Getz et al., 2012). In the case of afforestation, relevant knowledge will include confidence in and experience with growing trees, and knowledge of relevant markets, rules/regulations and programmes. These two types of knowledge are distinct in that landholders may have considerable experience of growing trees, gained when they planted trees for purposes other than afforestation, but little or no knowledge of afforestation specific issues such as carbon markets, and carbon trading rules/regulations.

2.5. Perceptions of attributes of afforestation

Landholder perceptions of the economic, environmental and land management attributes of new practices affect their willingness to adopt them, particularly the relative costs and benefits (relative advantage) of the new practice compared to current land uses (Pannell et al., 2006). The attributes of afforestation likely to be important to adoption go beyond simple economic return: for example, Gong et al. (2010), studying the world's first CDM afforestation project in China, found low rates of uptake despite

positive economic return for landholders, and argued this low had resulted from issues such as high risk and land tenure conflict. Coomes et al. (2008), meanwhile, identified that the opportunity costs and risks involved in afforestation were high compared to alternative land uses. The adoption literature suggests that perceptions of the economic, environmental and land management credibility of a new practice, and of its riskiness, complexity, flexibility and trialability, are important (Pannell et al., 2006).

We identified seven attributes of afforestation that may influence likelihood of adoption. The first of these is the economic costs and benefits of afforestation, including the overall financial return achieved from afforestation, perceived prices, and the impact of tree planting on property value (Jindal et al., 2012). The second is the impacts of afforestation on land management complexity (Pannell et al., 2006). Afforestation is often complex for landholders to adopt, as it involves practices that are unfamiliar and complex for landholders to learn (Mercer, 2004). The third is the impacts of afforestation on land management flexibility (Pannell et al., 2006). Multiple studies have identified that the length of time land must be dedicated to growing trees, and associated loss of land management flexibility, is a barrier to adoption of tree planting by farmers (e.g. Cacho et al., 2001; McDonagh et al., 2010). Similar findings have emerged for afforestation internationally, with concerns that planting trees for long time periods (typically a minimum 100 years) reduces the flexibility the landholder has to respond to changing land management circumstances (e.g. Jindal et al., 2012). Fourth, the on-property environmental costs and benefits of afforestation, such as impacts of trees on the incidence of weeds, pest animals, birds, and water quantity and quality, have been found to influence willingness to plant trees for environmental and commercial purposes (Schirmer et al., 2012). Fifth, perceptions of off-property environmental costs and benefits of afforestation, such as effectiveness at offsetting greenhouse gas emissions, and environmental benefits extending beyond the landholder's property, are likely to influence adoption (Schirmer et al., 2012). The sixth attribute is the risk that afforestation will fail. We identified three specific risks in our review of literature: the risk of trees not growing successfully; the risk of trees being killed in drought or fire; and the risk of failure of carbon markets due to market or governance failure, all issued identified in previous studies of afforestation (Coomes et al., 2008; Bozmoski and Hultman, 2010; Gong et al., 2010). Finally, the triability of afforestation may

Fig. 1. Hypothesised model of willingness to adopt afforestation.

influence landholder's willingness to adopt it. Multiple studies emphasise that the ease with which a landholder can trial a new practice will influence their willingness to adopt (Pannell et al., 2006), although no discussion of this was identified in the afforestation literature.

2.6. Model of afforestation adoption

We developed a theoretical model to identify how the factors identified above are likely to interact with each other, shown in Fig. 1. This model emphasises that willingness to adopt cannot be predicted by simply adding together the various factors identified as potentially influencing adoption, as some of these in fact determine each other. Instead, the various factors influence each other in a particular sequence as part of an overall learning process that leads to the adoption decision (Pannell et al., 2006). In this process, perceptions about whether afforestation has positive or negative attributes are formed. These perceptions are influenced by the design of afforestation, social acceptability of afforestation, socio-demographic attributes of landholders and their properties, and the landholder's skills, knowledge and experience of tree planting and afforestation. These factors have a range of effects on landholder perceptions regarding whether afforestation has positive or negative attributes—for example, if afforestation is generally considered socially unacceptable in a rural community, a much higher economic return may be required before landholders believe afforestation has positive economic attributes, a belief that in turn facilitates willingness to adopt. Similarly, perceptions about the risk of tree failure will depend on the landholder's knowledge of afforestation. The model suggests a need to analyse (i) which perceived attributes of afforestation best predict willingness to adopt, and (ii) which issues around social acceptability, skills and knowledge, and socio-demographic attributes best predict landholder's views about the attributes of afforestation. Further, it suggests the need to identify how this varies depending on the design of afforestation.

3. Case study and methods

Our study drew on data from a 2010 survey of rural landholders in the state of New South Wales, Australia. Tree planting for carbon

sequestration is in its relative infancy in Australia. A mandatory carbon trading scheme (the Carbon Pricing Mechanism) was introduced in 2012 (Keenan et al., 2012), following several years in which an active voluntary carbon market encouraged development of afforestation schemes. In 2010, prior to the introduction of the mandatory market, 24 businesses were involved in developing afforestation projects for the voluntary market in Australia (Dargusch et al., 2010). These firms used a number of mechanisms for establishing trees including share-farming, in which firms enter into a partnership with landholders to grow trees; leasing land from landholders; and outright purchase of land. In 2011 the Australian government passed its Carbon Farming Initiative (CFI), a legislative scheme for carbon offset crediting that provides a framework for regulating carbon credits in the farming and forestry sectors (Keenan et al., 2012). The CFI is intended to support trade in the voluntary market and to be compatible with any future mandatory market developed in Australia (Hug and Ahammad, 2011).

While large areas of land are potentially viable for afforestation in Australia - in 2008, an estimated 21.8 million ha of land was estimated to be economically viable for afforestation projects at a carbon price of $AU28 per tonne (Lawson et al., 2008) - the willingness of landholders to participate in afforestation projects is not well understood, similar to other countries (Dargusch et al., 2010; Bull and Thompson, 2011).

Our 2010 survey was conducted prior to the introduction of the CFI and subsequent mandatory market, but at a time when afforestation for the voluntary market was receiving widespread attention. The study region in the state of New South Wales extended from a high rainfall traditional forestry region in the east, in which a large area of tree plantations have been established and produce commercial timber products, to very low rainfall areas in the west (typically <400 mm annually) with little or no experience with commercial forestry activities (Fig. 2). In part of the study region, landholders had been offered options for planting trees for carbon sequestration for several years prior to 2010, although with a large area of land in the region very few landholders have undertaken afforestation. The region is dominated by relatively large cropping and grazing properties, typically between 100 and 2500 hectares in size (ABS, 2006), with both irrigated and dryland agriculture practiced.

Fig. 2. Location of study region, New South Wales, Australia.

The survey examined the factors that help and hinder rural landholders in adopting tree planting for carbon sequestration. Prior to design of the questionnaire, seven qualitative focus groups were held with 32 landholders to gather data to inform question design. A draft questionnaire was designed and tested with two landholders before revision. The final questionnaire was 12 pages in length and asked questions about characteristics of the landholders property; characteristics of the landholder (e.g. age, gender); the landholder's previous experience with tree planting; views about the positive and negative attributes of tree planting in general and tree planting for afforestation; values and beliefs about land management, climate change, the obligations of farmers, and perceived views of others about afforestation; willingness to adopt afforestation and how willingness varied for different designs of afforestation; barriers and incentives for adoption of afforestation; and level of knowledge about afforestation. Landholders were sampled through random selection from a database of contacts held by FarmBase, a commercial database that covers over 60% of Australian farming households, and which provided better coverage and more accurate addresses than alternative sources of contact information. Surveys were distributed to 1000 landholders living within the region identified in Fig. 2. A 40.0% response rate was achieved, with 352 valid survey responses from a final valid survey sample of 880 (120 of those who received the survey were ineligible to participate due to incorrect address, the resident having died or shifted, and were removed from the sample). Further detail on survey methods, including a copy of the full questionnaire, is provided in Schirmer and Bull (2011).

Survey data were analysed using the Statistical Package for Social Sciences (SPSS). The specific statistical analyses used are described when presented in the results.

Our results are presented in two sections. First, the likelihood of widespread adoption of afforestation in the case study region is examined for different design scenarios. Second, we analyse factors influencing willingness to adopt, first identifying significant bivariate relationships for each of the factors in Fig. 1, and using our results to respecify this model, before using a two-stage linear regression analysis to identify variables that explain the greatest amount of variance in willingness to adopt.

4. Likelihood of widespread adoption of afforestation

We examined the willingness of landholders to adopt afforestation in general, and under different design scenarios. Landholders were asked about their previous experience with, and interest in, adopting afforestation. Our analysis of adoption was limited to landholders' stated willingness to adopt, rather than actual rates of adoption. This is because too few of the respondents had adopted afforestation to enable a comparison of adopters and non-adopters: only 4.1% (14 landholders, n = 345) had planted trees specifically for carbon sequestration, and two of these were not willing to do so again in the future. While statements of intention do not always match subsequent behaviour (Sideridis, 2001), analysing willingness to adopt still provided a meaningful way to compare landholders who had different levels of initial interest in adoption, a prerequisite for subsequent adoption.

4.1. Willingness to adopt afforestation in general

When asked about their willingness to adopt afforestation in general, 19.2% of 345 landholders stated they had no interest in adopting afforestation at any time in the future (non-adopters), 67.2% indicated they might consider it in the future (possible adopters), and 13.6% were actively considering adoption at the time of completing the survey (10.1%), or had either already adopted and would do so again (3.6%) (likely adopters).

4.2. Design of afforestation

While landholders indicated a high general willingness to consider adopting, their willingness varied substantially depending on how afforestation was designed. We designed several realistic afforestation scenarios that specified factors such as the proportion of land to be planted, tree management, and type of land to be planted. The scenarios were developed based on review of current carbon sequestration planting programmes offered to landholders in Australia, including review of websites of four carbon tree planting organisations-CO2 Australia; Greening Australia; Carbon Neutral; and Carbon Conscious; and of the afforestation design factors likely to influence adoption more broadly, identified in the previous section. Landholders were then asked what level of financial return they would need to be offered to be willing to adopt carbon tree planting under each scenario. The full range of scenarios is reported in Schirmer and Bull (2011). Here, we focus on the three that best reflect actual carbon tree planting practice by landholders in Australia and internationally. Table 1 compares willingness to adopt under the three scenarios, and describes each.

Table 1 shows that, of the three scenarios, landholders were most willing to consider adopting small marginal plantings, in which trees are planted on a small proportion of their marginally productive land, and managed directly by the landholder; only 14.2% would never consider adopting small marginal plantings, while 27.1% said they would adopt them if they achieved the same return or less than the land makes now, and the remainder would require 10-20% more than current returns. Landholders were least willing to consider adopting large productive plantings, in which a large proportion of their more productive land is planted and managed by an external organisation (e.g. a carbon sequestration company); 66.0% would never do adopt this type of afforestation, and only 6.4% would do it for their current return or less. Meanwhile, 48.6% of landholders would never adopt large marginal plantings where much of their marginal land was planted by an outside organisation, while 15.5% would do it for their current return or less.

The variation in willingness to adopt between scenarios shows the importance of going beyond general measures of willingness to adopt to examine more realistic scenarios, although in all cases general willingness to adopt and willingness to adopt under the three scenarios was strongly and significantly related (Table 1).

Our analysis shows that while a large majority of landholders are willing to consider adopting afforestation, their willingness is not global: there is very low willingness to consider adopting the large-scale afforestation plantings that are common under the CDM and in the voluntary market, and a strong preference for plantings on small areas of non-productive land.

Landholders were then asked if changing a number of aspects of afforestation design would make them less likely to consider planting trees, more likely to, or neither more or less likely to. Landholders expressed strong preferences for afforestation that had the following design attributes: use of native species, planting of smaller rather than larger areas of land, use of marginal rather than highly productive land, shorter-term rather than longer-term plantings, and plantings that provide co-benefits for either livestock or the environment. In all but one case, the design preference was significantly related to a landholder's overall willingness to adopt (Table 2). Non-adopters were significantly less likely than possible or likely adopters to indicate they would be willing to adopt under any of the design scenarios presented. Likely adopters, and to a lesser extent possible adopters, were significantly more likely to indicate that they were likely to consider planting under all scenarios. This suggests that having an initial willingness to consider adoption results in greater likelihood

Table 1

Willingness to adopt carbon tree planting under three scenarios.

Scenario name and definition % of landholders who indicated the minimum return required for them to consider planting trees for carbon sequestration under this scenario was... Relationship to general willingness to adopt

...10% less than current return . . .equivalent to current return . . .10% more . . .20% more than current than current return return . . .not willing to do this for any amount (H, p, n)a

Small marginal (n = 325). Trees are planted on a small proportion of a landholder's marginal land, with the landholder providing all labour. Marginal land means land of low productivity for the agricultural 3.7% 23.4% 17.5% 41.2% 14.2% 17.9, <0.000, 321 "

enterprises the landholder undertakes, and is often in degraded condition landholder's 'marginal' (defined as low productive land, often subject to environmental degradation), and will support landholders

Large marginal (n = 323). Trees are 2.8% 12.7% 13.0% 22.9% 48.6% 26.6, <0.000,319""

planted on a large proportion of a landholder's marginal land, with an outside organisation providing labour and land management expertise. This is a common model, used by groups such as Carbon Conscious and Carbon Neutral, who prefer minimum planting sizes of 100 ha or more that often take up a large proportion of a property, but focus on marginal land

Large productive (n = 324). Trees are 1.2% 5.2% 7.7% 19.8% 66.0% 18.9, <0.000,321""

planted on a large proportion of a landholder's productive land, with an outside organisation providing labour and land management expertise. Groups that seek to plant large areas of land will often plant more productive land, and in some cases have minimum soil productivity standards to ensure particular minimum levels of tree growth and hence carbon

sequestration._

a Analysed using the Kruskal-Wallis H test. Significant relationships are indicated by bold text.

" p < 0.05.

"" p < 0.01.

of adoption under any afforestation design, and vice versa. It points to the importance of understanding the underlying factors that influence willingness to adopt, as well as understanding how the design of afforestation influences likelihood of adoption.

5. Factors influencing willingness to adopt

The second part of our analysis analysed which factors best explain landholder willingness to adopt. To do this, we drew on the theoretical model in Fig. 1. Our goal was to identify which of the various factors in the model were significant in explaining willingness to adopt. This required a two-stage analysis. First, a large number of variables that measured different aspects of each factor in the model were analysed using bivariate tests, to identify which had any relationship to willingness to adopt. Based on the results of the bivariate tests, we removed some variables which, while fitting the theoretical model, had no relationship to willingness to adopt. This resulted in a smaller, but still overly large number of significant variables. We undertook a two-stage stepwise regression to identify which of the remaining variables explained the greatest amount of variance in willingness to adopt.

In both the bivariate analyses and regression analysis, we included four dependent variables: general willingness to adopt,

and willingness to adopt under the small marginal, large marginal and large productive scenarios. The calculation of each of these four dependent variables is described in Table 3. Analysing four scenarios of willingness to adopt enabled us to compare the consistency of results of regression analyses across different afforestation scenarios. This cross-validation of results of the regressions improved the robustness of our analysis, with the different scenarios providing multiple overlapping samples (Sideridis, 2001).

5.1. Bivariate analyses

5.1.1. Social acceptability of afforestation

Landholders were asked the extent to which they agreed or disagreed with several statements regarding appropriate land use, climate change, farmer's responsibilities regarding environmental outcomes, and their beliefs about how others view afforestation. In bivariate analyses, almost all of the statements were significantly correlated with willingness to adopt, as shown in Table 4 (further detail is provided in Table S1). In particular, landholders were significantly less likely to be willing to adopt afforestation under all four scenarios if (i) they believed it was inappropriate to grow trees on good agricultural land; (ii) they were sceptical about climate

Afforestation design preferences and relationship to overall willingness to adopt.

% of all landholders who Mean score of different types of landholders (measured on Relationship to general

indicated changing scale of 1-5, 1 =less likely to consider planting, 5 = more willingness to adopt

afforestation design in likely) (H, p, n)a

this way would make them more likely to

consider adoption _

Non-adopters Possible adopters Likely adopters

Plant locally growing native tree 67.3% (n = 342) 3.6 3.9 4.4 14.9, 0.001, 337""

species 15.2, <0.000, 332 "

Plant high yielding tree species 42.7% (n = 337) 2.7 3.4 3.6

Landholder plants and manages trees 47.2% (n = 34l) 2.9 3.4 3.8 13.0, 0.002, 336 "

Outside organisation plants and 36.3% (n = 339) 2.2 3.1 3.0 24.3, <0.000, 334""

manages trees

Trees planted on marginal land 70.3% (n=334) 3.4 4.0 4.2 10.0, 0.007, 331""

Trees planted on productive land 4.2% (n=337) 1.4 1.7 1.8 8.8, 0.012, 332"

Trees planted on small proportion of 73.3% (n = 34l) 3.4 4.0 4.1 10.9, 0.004, 336""

property

Trees planted on large proportion of 3.3% (n = 341) 1.4 1.5 1.8 2.8, 0.251, 336

property

Trees providing environmental benefits 74.8% (n = 341) 3.4 4.1 4.5 29.9, <0.000, 336""

as well as carbon benefits

Trees provide stock shade and shelter 81.0% (n = 341) 3.6 4.2 4.4 21.7, <0.000, 336""

as well as carbon benefits

Trees grown for commercial wood 63.1% (n = 341) 3.0 3.9 3.7 14.0, 0.001, 336""

production as well as carbon

sequestration 30.6, <0.000, 334""

Trees grown for 30-40 years instead of 58.7% (n = 339) 2.8 3.8 4.0

currently required 100 years

a Analysed using the Kruskal-Wallis H test. Significant relationships are indicated by bold text. * p < 0.05. " p < 0.01.

change and associated science, although this relationship was weaker than that between views about the appropriateness of growing trees on agricultural land and adoption; (iii) they believed landholders shouldn't be asked to plant trees to fix climate problems caused by others; or (iv) they felt their neighbours or members of the broader community would disapprove if they adopted afforestation.

Supplementary material related to this article can be found, in the online version, at doi:10.1016/j.gloenvcha.2013.11.009.

Views about the land management obligations of farmers— whether farmers have a responsibility to manage land to provide benefits for the wider community, or prioritise economic return over achieving environmental outcomes-were largely uncorrelat-ed with willingness to adopt under any scenario.

5.1.2. Landholder skills, knowledge and experience relevant to afforestation

The large majority of respondents - 96.5% - had planted trees in the past for purposes other than carbon sequestration (e.g. shade and shelter for stock, to rehabilitate degraded land, or to improve how their property looked). However, only 6.2% had planted trees with the intent of achieving a commercial return in the form of carbon payments or wood production. Landholders thus had a high degree of knowledge about tree planting in general, but not of tree planting for commercial return.

Most landholders reported having low levels of knowledge about tree planting for carbon sequestration-particularly available programmes, prices and markets, and government rules and regulations. Relationships between level of knowledge about tree

Table 3

Definition and calculation of 'willingness to adopt' variables.

Willingness to adopt variable

Calculation of variables used in analyses

General WTA

Small marginal

Large marginal Large productive

Landholders categorised as falling on a linear spectrum from non-adopter to likely adopter using a 6-point scale (without specifying design of afforestation). Each category calculated as follows:

• Non-adopters were those who reported they have no interest in planting trees for carbon sequestration or who had previously adopted but would not do so again. If non-adopters also indicated they were not interested in planting trees for any purpose, they received a score of 1; if they indicated interest in planting more trees on their land in general, they received a score of 2.

• possible adopters were those who reported they might consider planting trees for carbon sequestration in the future (score of 3 or 4, based on whether they indicated they wished to plant more trees on their farm in general)

• likely adopters were those reporting they were actively considering planting trees for carbon sequestration in the future and those stating they have planted trees for carbon sequestration and would like to plant more. (score of 5 or 6, based on whether they indicated they wished to plant more trees on their farm in general)

For each landholder, calculated as the sum of the General WTA score and the amount the landholder would need to be paid to be willing to adopt afforestation under the 'small marginal' scenario. For the latter 1 = not willing to adopt for any payment, 2 = require 20% more than current return, 3 = require 10% more than current return, 4 = require current return, 5 = require 10% less than current return. This resulted in a score of between 2 and 11 for each landholder, which was used as a continuous dependent variable in subsequent analyses.

Sum of General WTA score and amount landholder would need to be paid to adopt afforestation under the 'large marginal' scenario (coding as specified in small marginal)

Sum of General WTA score and amount landholder would need to be paid to adopt afforestation under the 'large productive' scenario (coding as specified in small marginal)

Relationship between landholder values and beliefs and willingness to adopt.

Variable

Statement (respondents asked extent to which they disagreed or agreed on 5-point Likert scale)

% landholders who agreed or strongly agreed with statement

Relationship between statement about social acceptability and willingness to adopt in general and for three scenarios (rs; * =p < 0.05; * * = p < 0.01)a

General WTA Small

marginal

Large marginal Large

productive

Beliefs about appropriate use of agricultural land

GOOD USE 1 Good agricultural land should grow

food, not trees

GOOD USE 2 It is acceptable to grow trees on good

farm land

GOOD USE 3 In general, it is more acceptable to plant

trees on marginal land than good agricultural land

Beliefs about climate change

CC BELIEFS 1

CC BELIEFS 2

CC BELIEFS 3

I believe the science behind climate change is doubtful

Human use of fossil fuels is changing the climate

Farmers shouldn't be asked to plant trees to fix the climate problems caused by other people

Beliefs about farmer's land management obligations

OBLIGATION 1 Farmers have a responsibility to

manage their land to provide benefits for the wider community

OBLIGATION 2 Farmers have to prioritise making an economic return over improving environmental outcomes

OBLIGATION 3 Farmers should be paid to manage their land to provide benefits for the wider community

Social acceptability of afforestation in community

COMMUNITY 1 My neighbours would disapprove if I planted trees for carbon sequestration

COMMUNITY 2 Planting trees for carbon sequestration is viewed positively by people in my community

55.4% (n = 343) 39.8% (n = 347) 78.5% (n = 348)

61.1% (n = 319) 37.3% (n = 300) 42.1% (n = 328)

64.9% (n=345) 66.7% (n=336) 78.6% (n = 340)

8.7% (n=263) 31.9% (n = 238)

-0.288* 0.321*

- 0.009

- 0.147* 0.173*

-0.237*

0.060 -0.077 0.173*

-0.282* 0.368*

-0.308* 0.254* -0.016

-0.220* 0.170*' -0.265*

-0.165* 0.272*

-0.315* 0.335* -0.093

-0.137* 0.122* -0.205*

0.016 0.035

0.145*

-0.200* 0.245*

-0.326* 0.340* -0.191*'

-0.160*' 0.137* -0.201*'

0.033 -0.079 0.144*'

-0.204* 0.288*

Analysed using the Spearmans Rank Order Correlation (two-tailed). See Schirmer and Bull (2011 a See Supplementary data for further details of the analysis including all p-values and sample

) for detailed presentation of descriptive statistics for these statements. sizes for each bivariate analysis.

Table 5

Relationship between landholder skills, knowledge and experience and willingness to adopt.

Variable

Statement

Percentage of respondents, n

Relationship between statement about skills and experience and willingness to adopt in general and for three scenarios (rs, p, n)a

General

Small marginal

Large marginal

Large productive

Previous experience with tree planting

TREE SKILLS 1 % of landholders who have planted

trees for any purpose in the past TREE SKILLS 2 % landholders who have planted trees

for commercial return in the past TREE SKILLS 3 % landholders reporting most trees

planted in past have grown successfully

Self-rated knowledge of afforestation

KNOWLEDGE 1 % with good or very good knowledge of current programmes that pay landholders to grow trees on their land for carbon sequestration KNOWLEDGE 2 % with good/very good knowledge of government rules and regulations regarding carbon tree planting KNOWLEDGE 3 % with good/very good knowledge of

carbon prices and markets KNOWLEDGE 4 % with good/very good knowledge of where to find information on planting trees for carbon sequestration

96.5% (n = 346) 6.2% (n = 346) 59.6% (n = 322)

5.4% (n = 347)

3.5% (n = 347)

2.3% (n = 345) 7.3% (n = 343)

Not tested as small sample with no experience means results not robust Not tested as small sample with experience means results not robust -0.074 -0.032 -0.008 -0.015

-0.116

-0.149

-0.177

-0.162

-0.195 -0.194

-0.138* -0.097

-0.200**

-0.244**

-0.249**

- 0.106

0.181**

0.236**

0.218** 0.118*

Analysed using the Spearmans Rank Order Correlation (two-tailed). Significant relationships are indicated by bold text. See Schirmer and Bull (2011) for detailed presentation of descriptive statistics for these statements. a See Supplementary data for further details of the analysis including all p-values and sample sizes for each bivariate analysis.

* 0.05 level.

** 0.01 level.

Relationship between landholder perceptions of afforestation attributes and willingness to adopt.

Variable

Statement

% who agreed/strongly agreed with statement

Relationship between statement about social acceptability and willingness to adoptb

General

Small marginal Large marginal Large productive

Perceived economic costs and benefits of afforestation

ECON 1 Current carbon tree planting programmes

don't offer enough money to be worth taking up

ECON 2 Planting trees for CSa would increase the

value of my property ECON 3 Planting trees for CS would benefit me by

providing a diversified income stream ECON 4 Future carbon prices are likely to be high

enough for me to consider CS

Land management complexity

COMPLEX 1 Planting trees for CS would make my farm management more complicated

Land management flexibility

FLEX 1 The long timeframes involved reduce my

willingness to consider growing trees for CS FLEX 2 Planting trees for CS would reduce my

ability to manage my land flexibly

Off-property environmental costs and benefits

OFF-PROP 1 It is important that everyone plants trees on their land to contribute to reducing climate change

OFF-PROP 2 Planting trees on one property is too little to

make any difference to climate change OFF-PROP 3 Planting trees has environmental benefits beyond my property

On-property environmental costs and benefits

ON-PROP 1 Planting trees has environmental benefits

on my property ON-PROP 2 Planting trees improves how my property looks

Risk of tree failure

TREE RISK 1 It is too risky to plant trees for CS as they

might be killed by drought or fire TREE RISK 2 I have seen too many failed attempts at tree planting to feel confident in planting trees on my property for CS

Risk of market failure

MKT RISK 1 I feel confident there will be a long term

market for carbon stored in trees MKT RISK 2 There is a high risk that carbon markets will disappear in a few years time

Trialability

TRIAL 1 More local trials are needed before I would

feel confident planting trees for CS

73.9% (n=222)

21.9% (n = 242) 40.1% (n = 247) 52.5% (n = 198)

34.2% (n=286)

51.0% (n = 304) 38.0% (n = 305)

39.8% (n = 321)

37.0% (n = 308) 84.0% (n = 332)

90.9% (n=342) 92.3% (n = 349)

25.6% (n = 313) 15.9% (n = 308)

35.7% (n = 213) 61.0% (n = 228)

54.2% (n = 319)

-0.072

0.460 0.416 0.414

0.170 0.156

-0.112

0.313 0.285 0.304

-0.211 -0.228

-0.237 -0.151 -0.307 -0.273

-0.318 -0.147

0.171 0.116

-0.300 -0.290 -0.149 -0.168

0.309 0.353

-0.280 -0.333

-0.117 -0.101

-0.080

0.295 0.426 0.331

-0.223

-0.143 -0.227

-0.206 0.204

0.151 0.188

-0.219 -0.149

0.274 -0.242

-0.006

-0.088

0.314 0.387 0.347

-0.241

-0.131 -0.283

-0.179 0.143

0.086 0.133

-0.236 -0.175

0.258 -0.236

-0.050

a CS = carbon sequestration. Analysed using the Spearmans Rank Order Correlation (two-tailed). Analysed using the Spearmans Rank Order Correlation (two-tailed). b See Supplementary data for further details of the analysis including all p-values and sample sizes for each bivariate analysis. Significant relationships are indicated by bold text. See Schirmer and Bull (2011) for detailed presentation of descriptive statistics for these statements. 0.05 level. 0.01 level.

planting for carbon sequestration and adoption were somewhat complex (Table 5). Those who reported having high levels of knowledge were significantly less likely to be willing to adopt afforestation. This finding is difficult to interpret, as it is based on a landholder's perception of their level of knowledge: it is equally possible that landholders with good knowledge of afforestation used the information they obtained to decide afforestation was not an attractive proposition; or that landholders who were not willing to adopt felt they had high knowledge despite not having accessed large amounts of information.

5.1.3. Attributes of afforestation

Landholders were significantly more likely to adopt if they believed afforestation had one or more of the following five

attributes (Table 6): (i) economic benefit in the form of increased property values, diversified income, or high prices; (ii) low complexity; (iii) off-property environmental benefits; (iv) on-property environmental benefits, although this did not hold for all scenarios and there were generally lower significance and effect sizes than for other relationships; and (v) low risk related to failure of trees to grow or failure or markets. Landholder views about the need for more trials of afforestation, however, were not typically significantly related to willingness to adopt.

5.1.4. Characteristics of the landholder and their property

While emphasised in much of the adoption literature, neither the socio-demographic characteristics of landholders, or the type of property they managed, were typically significantly related to

Relationship between landholder and property characteristics and willingness to adopt.

Variable Landholder/property attribute and measurement unit Relationship between statement about social accept-

ability and willingness to adoptc

General Small Large Large

marginal marginal productive

Property characteristics

PROP 1b Primary purpose of land (92.9% agriculture; 7.1% other, n = 352) 0.185 1.465 0.396 0.738

PROP 2a Area of land (hectares) -0.036 -0.092 -0.116* -0.093

PROP 3b Type of agriculture (pure grazing, mixed grazing/cropping, pure cropping, horticulture) 0.733 1.426 2.381 2.768

PROP 4a % marginal land (percent of total property considered marginal for agriculture) 0.035 0.109 0.076 0.020

PROP 5a Length of property ownership (years) -0.163" -0.149"" -0.239" -0.205**

PROP 6b Whether have conservation covenant (yes/no) 1.474 3.241 1.374 2.596

Landholder characteristics

LHDR 1 Gender (male 83.7%, female 16.3%, n = 349) 0.579 7.381" 1.118 0.078

LHDR 2a Age (years) -0.155" -0.092 -0.109 -0.084

LHDR 3b Main occupation (farmer 84.9%, non-farmer 15.1%, n = 352) 1.700 1.722 4.758* 2.304

LHDR 4a Off-farm income (% of total household income) 0.097 0.103 0.091 0.067

LHDR 5a Generations of family who have been involved in farming (number) 0.016 0.036 -0.084 -0.057

LHDR 6b Succession plans (sell outside family, sell to family, hand down within family, unsure) 3.335 0.937 3.244 3.994

LHDR 7a Formal education (highest level achieved, from primary school to tertiary education) 0.055 0.045 0.077 0.032

LHDR 8a Income (household income 2009-10, $) -0.073 -0.020 -0.054 -0.032

a Analysed using the Spearmans Rank Order Correlation (two-tailed). b Analysed using the Kruskal-Wallis H test.

c See Supplementary data for further details of the analysis including all p-values and sample sizes for each bivariate analysis. Significant relationships are indicated by bold text. See Schirmer and Bull (2011) for detailed presentation of descriptive statistics. * 0.05 level. " 0.01 level.

willingness to adopt (Table 7). No significant relationship was identified in bivariate analyses of willingness to adopt and primary purpose of land management, area of land managed, type of agriculture undertaken, the proportion of marginal land on a property, whether a landholder had a conservation covenant on part of their land, income, education, succession plans, or history of involvement in agriculture. Gender, age and occupation were each significantly related to willingness to adopt in only one of the adoption scenarios examined, and no theoretical argument was identified to explain why a significant relationship existed for one scenario and not others.

The only attribute consistently and significantly related to willingness to adopt was length of property ownership: landholders who had owned their property for a shorter time were more likely to be willing to adopt than those who had owned their property longer.

5.2. Regression analysis: explaining variation in willingness to adopt

Large numbers of significant bivariate relationships were identified, all consistent with our originally identified theoretical arguments. We undertook regression analysis to identify the subset of variables that explained the most variation in willingness to adopt. First, we re-specified our theoretical model, and populated it with the variables identified as having most significance in bivariate analyses. The revised model is shown in Fig. 3. We then tested this model using a two-stage stepwise linear regression. Two stages were necessary to reflect the learning process in which (i) perceived attributes of afforestation influence willingness to adopt, and (ii) social acceptability, knowledge and landholder characteristics influence a landholder's perceptions about attributes of afforestation. While we considered using hierarchical regression, we chose to undertake separate linear

Fig. 3. Respecified model tested in linear regression analysis.

Table 8

Predictors of willingness to adopt (afforestation attributes model).

Final model B Std Err B b Sig. R2

General WTA <0.000 0.38

Constant 1.086 0.364

ECON 3 0.147 0.055 0.214 0.008

FLEX 2 -0.136 0.043 -0.228 0.002

ON-PROP 2 -.158 0.057 0.190 0.006

TREE RISK 1 -0.167 0.055 -0.245 0.003

TREE RISK 2 0.127 0.056 0.182 0.024

ECON 2 0.110 0.053 0.162 0.040

Small marginal <0.000 0.25

Constant 2.836 0.717

ECON 3 0.457 0.116 0.291 <0.000

FLEX 2 -0.323 0.100 -0.235 0.002

ON-PROP 1 0.290 0.118 0.179 0.015

Large marginal <0.000 0.22

Constant 1.376 0.386

ECON 3 0.784 0.119 0.464 <0.000

Large productive <0.000 0.18

Constant 2.457 0.500

ECON 3 0.492 0.14 0.355 <0.000

FLEX 2 -0.181 0.091 -0.149 0.049

* p < 0.05. " p < 0.01.

regressions for each model component, in order to better be able to present and explore significant predictors in each component.

First, we identified which perceived attributes of afforestation best explained variation in willingness to adopt under the four scenarios (general WTA, small marginal, large marginal and large productive). Results are shown in Table 8. Of the various attributes of afforestation, those that explained most variation in willingness to adopt across the different scenarios were the extent to which a landholder believed that afforestation would (i) provide economic benefit through a diversified income stream (positive related to willingness to adopt), and (ii) reduce land management flexibility (negatively related to willingness to adopt). If a landholder believed planting trees provided environmental benefit on their property, they were more likely to be willing to adopt under the general WTA and small marginal scenarios, but not under the large marginal or large productive scenarios.

To test the other part of the model - the hypothesised influence of the social acceptability of afforestation, landholder demographics, and knowledge about afforestation on the opinions

landholders form about attributes of afforestation - we selected as our dependent variables the two attributes of afforestation most commonly linked to willingness to adopt: (i) perceived economic benefits of afforestation (income diversity) and (ii) perceived impact of afforestation in flexibility of land management (Table 9). We also regressed variables related to social acceptability, demographics and knowledge against general willingness to adopt, as, although mediated by perceptions of attributes of afforestation, these variables should be significant predictors of willingness to adopt in their own right.

In all three regressions, similar variables explained a significant proportion of variance in willingness to adopt, perceptions about income diversity and perceptions about land management flexibility respectively. If a landholder believed it was acceptable to grow trees on good agricultural land, or that others in their community view afforestation positively, they were significantly more likely to have positive views about the impacts of afforestation. Landholder's self-rated level of knowledge of carbon prices and markets and, in the case of land management flexibility,

Table 9

Predictors of attributes of afforestation and willingness to adopt—social acceptability and knowledge model.

Final model B Std Err B b Sig. R2

Dependent variable: ECON2 <0.000 0.25

Constant 1.174 0.212

GOOD USE 2 0.137 0.038 0.259 <0.000

COMMUNITY 2 0.219 0.051 0.304 <0.000

KNOWLEDGE 3 -0.154 0.050 -0.220 0.003

Dependent variable: FLEX 2 <0.000 0.37

Constant 2.343 0.365

CC BELIEFS 3 -0.228 0.054 -0.302 <0.000

COMMUNITY 2 0.344 0.070 0.335 <0.000

GOOD USE 2 -.186 0.051 0.255 <0.000

Dependent variable: WTA general <0.000 0.25

Constant 1.174 0.212

GOOD USE 2 0.137 0.038 0.259 <0.000

COMMUNITY 2 0.219 0.051 0.304 <0.000

KNOWLEDGE 3 -0.154 0.050 -0.220 0.003

* p < 0.05. " p < 0.01.

their beliefs about whether landholders should be asked to plant trees to solve climate change problems caused by others, also explained a significant proportion of variance.

6. Discussion

Our contention in this paper is that in many countries, afforestation can only make a significant contribution to climate change mitigation efforts if private landholders are willing to adopt it; and that, while important, estimates of the area of land on which afforestation would theoretically be biophysically and economically viable are not adequate predictors of actual adoption. This reflects growing recognition that adoption of afforestation is not determined solely by biophysical and economic factors (Polglase et al., 2013). Our results support these hypotheses, and suggest some important issues need to be addressed to achieve large-scale uptake of afforestation, both in Australia and internationally.

First, our results show landholders are not opposed to afforestation, as long as it does not present a threat to their current agricultural activities or values about land management. Landholders strongly prefer the idea of small scale tree plantings on their marginally productive land, and are much less likely to be willing to consider adopting (let alone actually adopt) larger scale afforestation, particularly on their productive land. This suggests that landholders seek to place afforestation at the margins of their enterprise, where it will not interfere with existing agricultural production or views about acceptable use of agricultural land, and where afforestation thus presents less economic or social risk to the landholder. This is consistent with findings of other studies of afforestation; in Africa, Reynolds (2012 identified that opportunity costs of planting trees on good agricultural land were so high that smaller plantings on lower quality land were preferable for landholders despite having lower economic return. This finding suggests landholders will accept afforestation only if its design enables them to minimise the opportunity cost of afforestation, and to continue their existing land management activities and socio-economic relations with little disturbance. The preferences we identified for small-scale plantings on marginal land are likely to be relevant beyond our case study region. As identified earlier in this paper, small-scale afforestation on marginal land appears more socially acceptable than large-scale plantings on more productive land in multiple countries (e.g. Schirmer, 2007). It is unsurprising that the same finding applies to afforestation, and it is likely to apply in many regions worldwide.

Is it a problem if landholders are only willing to adopt small-scale plantings on marginal land? Some recent studies suggest that afforestation may be most economically viable on marginal land, or in designs such as row plantings, as these designs act to reduce the opportunity cost to landholders of adopting afforestation (Torres et al., 2010; Paul et al., 2013). The greatest challenge is not the placement of trees on marginal land, but the preference for small-scale plantings: small-scale plantings incur higher transaction and monitoring costs, and hence are higher cost afforestation options per unit of carbon sequestered (Smith and Scherr, 2003; Anderson and Zerriffi, 2012). Monitoring and verifying carbon sequestration in many small plantings dispersed across a large area is more costly than for large single areas of afforestation, as it increases the number of contracts that must be negotiated and executed, and the number of properties and individual sites to be monitored; previous studies suggest that establishing multiple small plantings may render afforestation projects uneconomic by creating a 'transaction costs trap' (van Kooten et al., 2002). This is particularly concerning given that other studies suggest that low rates of uptake of afforestation under the CDM are at least in part a

result of high transactions costs, particularly administrative burden (Tal and Gordon, 2010; Thomas et al., 2010). In addition, the marginal land preferred for planting by landholders will often yield lower tree growth rates compared to their more productive land, and hence lower rates of carbon sequestration and reduced mitigation success (Reynolds, 2012). Thus modifying afforestation to make it more attractive to landholders has potential to reduce its effectiveness for climate change mitigation, if not designed appropriately. The challenges presented to those considering design of afforestation is thus to balance the competing demands of desire for multiple small plantings, which will increase total uptake of afforestation, versus the complexity of carbon management, measurement and verification in these plantings (Law and Harmon, 2011; Lee et al., 2013).

The predominance of larger-scale afforestation projects in the CDM (Razak et al., 2009) likely reflects the relatively lower transaction costs of larger-scale afforestation plantings. Our findings suggest that there will be a finite limit to land availability for larger-scale plantings, beyond which additional adoption will require more flexible, smaller scale plantings on marginal land; this is consistent with other studies (Torres et al., 2010). Smaller-scale plantings have evidence of success in some locations: in Africa Reynolds (2012) found small afforestation plantings on marginal land were often successful, despite having poorer economies of scale than larger plantings.

Our analysis highlights the importance of understanding not only landholder's surface beliefs about the attributes of afforestation, but understanding how those attributes are formed. Unsurprisingly, we identified that landholder views about the impacts of afforestation on their income stream, and their land management flexibility, were strongly significant in predicting willingness to adopt across almost all afforestation scenarios. The perceived environmental benefits of afforestation, meanwhile, were only meaningful predictors in the scenario where small marginal areas of land were to be planted, suggesting that where afforestation does not replace existing agriculture, it is more likely to be adopted for its noncommercial benefits, whereas locating it on land used for production will only occur if it is commercially viable. This adds to the ongoing debate about the effect of afforestation on broader environmental characteristics, such as biodiversity, and supports the argument that afforestation should be explicitly designed to achieve broader environmental outcomes beyond carbon sequestration (Pawson et al., 2013).

When we turned to identifying which factors most strongly predicted landholder beliefs about the attributes of afforestation (i.e. its impacts on income, and flexibility of land management), however, the importance of underlying values and beliefs became apparent. While knowledge of afforestation markets and prices played a role, the most consistent predictors were landholder's beliefs about the social acceptability of afforestation in their community, and their beliefs about whether it is acceptable to grow trees on good agricultural land. This highlights the important role of agrarian ideals and values in the formation of beliefs about whether afforestation has positive or negative attributes and is, thus, worthwhile adopting. The agrarian ideals of most importance were about the imperative of food production, rather than landholder perceptions about whether human activities are contributing to climatic change, or the validity of climate change science. This finding is consistent with multiple studies that have highlighted the importance of productivist values to farmer identity internationally: farmers in many countries have a strong ethic which sees production of food as a moral good (Burton and Wilson, 2006). Our results suggest that afforestation challenges this productivist ethic, with the action of growing trees for carbon

not viewed as morally equivalent to growing food. While landholders had a strong stewardship ethic, and viewed the action of planting trees positively in terms of its environmental benefits both on- and off-property, their stewardship values only contributed significantly to willingness to adopt afforestation if the afforestation was on marginal land where it did not interfere with food production.

Overall, our results suggest that to achieve widespread adoption of afforestation, it is necessary to either develop afforestation designs that do not challenge existing values around the primacy of food production on agricultural land, or to attempt to shift deeply held productivist values. Given that landholders are willing to adopt afforestation on their marginal land for lower economic return compared to their productive land, and interested in planting for afforestation if it provides other co-benefits such as environmental improvement, models of afforestation that involve low economic return but target marginal land and provide co-benefits are likely to achieve adoption by a reasonable proportion of landholders. These must, however, have low transaction costs if they are to succeed. More broadly, our results support the increasing call for afforestation to be designed in ways that explicitly maximise socioeconomic benefits to landholders: designing AR in this way is essential not only from a moral point of view, with an obligation to ensure that afforestation reaches 'a standard of social benefits' if it is to form a significant part of the CDM (Boyd et al., 2007, p. 420), but also to achieve more widespread adoption of afforestation.

7. Conclusions

Afforestation/reforestation (afforestation) projects are a commonly promoted climate change mitigation strategy. Afforestation is often targeted at the private landholders who, in many countries, control the majority of the cleared agricultural land that is eligible for afforestation. Achieving uptake of afforestation requires convincing these landholders to adopt it on their land. Our study of landholder willingness to adopt afforestation in Australia demonstrates that the willingness of landholders to adopt afforestation depends on how that afforestation is designed. In particular, landholders prefer small tree plantings on less productive land, which minimise the disturbance afforestation presents to both land management, and landholder values about appropriate uses of agricultural land. Landholders are less willing to consider afforestation if it involves planting large areas, particularly large areas of their more productive land. Our results are consistent with findings on landholder preferences regarding afforestation in several other countries. Our results highlight that landholder's strongly held views about the importance of using agricultural land for food production present limitations to the extent of likely adoption of afforestation. Adoption is, however, highly likely when farmers are able to plant marginal land, and where plantings have co-benefits for the environment, even though this is likely to involve lower economic return compared to plantings on more productive land.

Acknowledgements

We acknowledge the time and effort provided by the many landholders who participated in focus groups and surveys for this project, and thank them for their contributions. This research was funded by (i) the Cooperative Research Centre for Forestry, Hobart; and (ii) a Science and Innovation Award from the Australian government Department of Agriculture, Fisheries and Forestry, with the award funded by Forest and Wood Products Australia. The funders had no involvement in project design, data collection, or

data analysis and interpretation. The research was approved by the Human Research Ethics Committee of the Australian National University (protocol 2010/224).

References

ABS (Australian Bureau of Statistics), 2006. Agricultural Census. Customised Data Order Supplied by ABS. Australian Bureau of Statistics, Canberra.

Anderson, E.K., Zerriffi, H., 2012. Seeing the trees for the carbon: agroforestry for development and carbon mitigation. Climatic Change 115 (3-4) 741-757.

Barlow, K., Cocklin, C., 2003. Reconstructing rurality and community: plantation forestry in Victoria, Australia. Journal of Rural Studies 19, 503-519.

Baumgart-Getz, A., Prokopy, L.S., Floress, K., 2012. Why farmers adopt best management practice in the United States: a meta-analysis of the adoption literature. Journal of Environmental Management 96 (1) 17-25.

Benitez, P.C., McCallum, I., Obersteiner, M., Yamagata, Y., 2007. Global potential for carbon sequestration: geographical distribution, country risk and policy implications. Ecological Economics 60, 572-583.

Blennow, K., Persson, J., 2009. Climate change: motivation for taking measure to adapt. Global Environmental Change 19, 100-104.

Boyd, E., May, P., Chang, M., Veiga, F.C., 2007. Exploring socioeconomic impacts of forest based mitigation projects: lessons from Brazil and Bolivia. Environmental Science and Policy 10 (5) 419-433.

Bozmoski, A.S., Hultman, N.E., 2010. Participant perceptions of risk and benefit in carbon forestry: evidence from Central Tanzania. The Journal of Environment and Development 19 (4) 4-27.

Bull, L., Thompson, D., 2011. Developing forest sinks in Australia and the United States—a forest owner's prerogative. Forest Policy and Economics 13,311-317.

Burton, R.J.F., Wilson, G.A., 2006. Injecting social psychology theory into conceptualisations of agricultural agency: towards a post-productivist farmer self-identity? Journal of Rural Studies 22 (1) 95-155.

Cacho, O., Greiner, R., Fulloon, L., 2001. An economic analysis of farm forestry as a means of controlling dryland salinity. The Australian Journal of Agricultural and Resource Economics 45 (2) 233-256.

Cacho, O., Hean, R., Ginoga, K., Wise, R., Djaenudin, D., Lugina, M., Wulan, Y., Subarudi, Lusiana, B., van Noordwijk, M., Khasanah, N., 2008. Economic potential of land-use change and forestry for carbon sequestration and poverty reduction, Technical Report No. 68. Australian Centre for International Agricultural Research, Canberra.

Cannell, M.G.R., 2003. Carbon sequestration and biomass energy offset: theoretical, potential and achievable capacities globally, in Europe and the UK. Biomass and Bioenergy 24, 97-116.

Ciscell, G.M., 2010. Beyond CO2lonialism: the potential for fair trade certification to embrace voluntary carbon offsets. International Journal of Sustainable Society 2 (2) 121-132.

Coomes, O.T., Grimard, F., Potvin, C., Sima, P., 2008. The fate of the tropical forest: carbon or cattle? Ecological Economics 65 (2) 207-212.

Corbera, E., Brown, K., 2010. Offsetting benefits? Analyzing access to forest carbon. Environment and Planning A 42 (7) 1739-1761.

Cossalter, C., Pye-Smith, C., 2003. Fast-wood Forestry: Myths and Realities. Centre for International Forestry Research, Bogor.

Dargusch, P., Harrison, S., Herbohn, J., 2010. How carbon markets have influenced change in the Australian forest industries. Australian Forestry 73, 165-170.

Diaz, D., Hamilton, K., Johnson, E., 2011. State of the Forest Carbon Markets 2011: From Canopy to Currency. Ecosystem Marketplace Report. Forest Trends, Washington, DC.

FAO (Food and Agriculture Organization of the United Nations), 2010. Global Forest Resources Assessment 2010: Main Report. FAO, Rome.

Flick, B., Caltabiano, N., Bentrupperbaumer, J., 2010. Retired farmers' perceptions regarding the ripairan forests in the Great Barrier Reef Lagoon catchment. Psychology and Society 3 (1) 107-126.

Gerber, J.-F., 2011. Conflicts over industrial tree plantations in the South: who, how and why? Global Environmental Change 21, 165-176.

Gong, Y., Bull, G., Baylis, K., 2010. Participation in the world's first clean development mechanism forest project: the role of property rights, social capital and contractual rules. Ecological Economics 69 (6) 1292-1302.

Hamilton, K., Chokkalingam, U., Bendana, M., 2010. State of the Forest Carbon Markets 2009: Taking Root and Branching Out. Ecosystem Marketplace Report. Forest Trends, Washington, DC.

Hug, B., Ahammad, H., 2011. The Economics of Australian Agriculture's Participation in Carbon Offset Markets. Outlook 2011. Australian Bureau of Agricultural and Resource Economics and Sciences, Australian Government, Canberra.

Jindal, R., Swallow, B., Kerr, J., 2008. Forestry-based carbon sequestration projects in Africa: potential benefits and challenges. Natural Resources Forum 32,116-130.

Jindal, E., Kerr, J.M., Carter, S., 2012. Reducing poverty through carbon forestry? Impacts of the N'hambita Community Carbon Project in Mozambique. World Development 40 (10) 2123-2135.

Kallio, A.M.I., Salminen, O., Sievanen, R., in press. Sequester or substitute—consequences of increased production of wood based energy on the carbon balance in Finland. Journal of Forest Economics (in press), Available online 5 July 2013.

Keenan, R.J., Caripis, L., Foerster, A., Godden, L., Peel, J., 2012. Science and the governance of Australia's climate regime. Nature Climate Change 2 (7) 477-478.

Knowler, D., Bradshaw, B., 2007. Farmers' adoption of conservation agriculture: a review and synthesis of recent research. Food Policy 32 (1) 25-48.

Law, B.E., Harmon, M.E., 2011. Forest sector carbon management, measurement and verification, and discussion of policy related to climate change. Carbon Management 2(1) 73-84.

Lawson, K., Burns, K., Low, K., Heyhoe, E., Ahammad, H., 2008. Analysing the Economic Potential of Forestry for Carbon Sequestration Under Alternative Carbon Price Paths. Australian Bureau of Agricultural and Resource Economics, Canberra.

Lee, C.M., Lazarus, M., Smith, G.R., Todd, K., Weitz, M., 2013. A ton is not always a ton: a road-test of landfill, manure, and afforestation/reforestation offset protocols in the US carbon market. Environmental Science and Policy 33, 53-62.

McDonagh, J., Farrell, M., Mahon, M., Ryan, M., 2010. New opportunities and cautionary steps? Farmers, forestry and rural development in Ireland. European Countryside 2 (4) 236-251.

McGinty, M.M., Swisher, M.E., Alavalapati, J., 2008. Agroforestry adoption and maintenance: self-efficacy, attitudes and socio-economic factors. Agroforestry Systems 73, 99-108.

Mercer, D.E., 2004. Adoption of agroforestry innovations in the tropics: a review. Agroforestry Systems 61 (1) 311-328.

Murray, B.C., McCarl, B.A., Lee, H-.C., 2004. Estimating leakage from forest carbon sequestration programs. Land Economics 80, 109-124.

Nilsson, S., Schopfhauser, W., 1995. The carbon-sequestration potential of a global afforestation program. Climatic Change 30, 267-293.

Ommani, A.R., Chizari, M., Salmanzadeh, C., Faraj Allah Hosaini, J., 2009. Predicting adoption behaviour of farmers regarding on-farm sustainable water resources management (SWRM): comparison of models. Journal of Sustainable Agriculture 33, 595-616.

Palmer, C., Silber, T., 2012. Trade-offs between carbon sequestration and rural incomes in the N'hambita Community Carbon Project, Mozambique. Land Use Policy 29 (1) 83-93.

Pannell, D.J., Marshall, G., Barr, N., Curtis, A., Vanclay, F., Wilkinson, R., 2006. Understanding and promoting adoption of conservation practices by rural landholders. Australian Journal of Experimental Agriculture 46, 1407-1424.

Paul, K.I., Reeson, A., Polglase, P., Crossman, N., Freudenberger, D., Hawkins, C., 2013. Economic and employment implications of a carbon market for integrated farm forestry and biodiverse environmental plantings. Land Use Policy 30(1) 496-506.

Pawson, S.M., Brin, A., Brockerhoff, E.G., Lamb, D., Payn, T.W., Paquette, A., Parrotta, J.A., 2013. Plantation forests, climate change and biodiversity. Biodiversity and Conservation 1-25.

Peters-Stanley, M., Hamilton, K., 2012. Developing Dimension: State of the Voluntary Carbon Markets 2012. Ecosystem Marketplace. Forest Trends, Washington, DC.

Polglase, P.J., Reeson, A., Hawkins, C.S., Paul, K.I., Siggins, A.W., Turner, J., Crawford, D.F., Jovanovic, T., Hobbs, T.J., Opie, K., Carwardine, J., Almeida, A., 2013. Potential for forest carbon plantings to offset greenhouse emissions in Australia: economics and constraints to implementation. Climatic Change 1-15.

Prokopy, L.S., Floress, K., Klotthor-Weinkauf, D., Baumgart-Getz, A., 2008. Determinants of agricultural best management practice adoption: evidence from the literature. Journal of Soil and Water Conservation 63, 300-311.

Race, D., Curtis, A., 2007. Adoption of farm forestry in Victoria: linking policy with practice. Australasian Journal of Environmental Management 14, 166-178.

Razak, S.A., Son, Y., Lee, W.-K., Cho, Y., Noh, N.J., 2009. Afforestation and reforestation with the clean development mechanism: potentials, problems, and future directions. Forest Science and Technology 5 (2) 45-56.

Reynolds, T.W., 2012. Institutional determinants of success among forestry-based carbon sequestration projects in Sub-Saharan Africa. World Development 40 (3) 542-554.

Richards, K.R., Stokes, C., 2004. A review of forest carbon sequestration cost studies: a dozen years of research. Climatic Change 63, 1-48.

Rootzen, J.M., Berndes, G., Ravindranath, N.H., Somashekar, H.I., Murthy, I.K., Sudha, P., Ostwald, M., 2010. Carbon sequestration versus bioenergy: a case study from South India exploring the relative land-use efficiency of two options for climate change mitigation. Biomass and Bioenergy 34 (1) 116-123.

Sathaye, J.A., Andrasko, K., 2007. Special issue on estimation of baselines and leakage in carbon mitigation forestry projects. Mitigation and Adaptation Strategies for Global Change 12, 963-970.

Schirmer, J., 2007. Plantations and social conflict: exploring the differences between small-scale and large-scale plantation forestry. Small-Scale Forestry 6, 19-33.

Schirmer, J., Bull, L., 2011. Planting Trees for Carbon Sequestration: What do Landholders Think? Cooperative Research Centre for Forestry, Hobart.

Schirmer, J., Clayton, H., Sherren, K., 2012. Reversing scattered tree decline on farms: implications of landholder perceptions and practice in the Lachlan catchment, New South Wales. Australasian Journal of Environmental Management 19 (2) 91-107.

Shaikh, S.L., Sun, L., van Kooten, G.C., 2007. Are agricultural values a reliable guide in determining landowners' decisions to create forest carbon sinks? Canadian Journal of Agricultural Economics 55, 97-114.

Sideridis, G.D., 2001. The causal role of goal importance for the explanation of student study behaviour: cross-validation with multiple samples. Educational Psychology 21 (3) 277-298.

Smith, J., Scherr, S.J., 2003. Capturing the value of forest carbon for local livelihoods. World Development 31 (12) 2143-2160.

Tal, A., Gordon, J., 2010. Carbon cautious: Israel's afforestation experience and approach to sequestration. Small-Scale Forestry 9 (4) 409-428.

Thomas, S., Dargusch, P., Harrison, S., Herbohn, J., 2010. Why are there so few afforestation and reforestation clean development mechanism projects? Land Use Policy 27, 880-887.

Torres, A.B., Marchant, R., Lovett, J.C., Smart, J.C.R., Tipper, R., 2010. Analysis of the carbon sequestration costs of afforestation and reforestation agroforestry practices and the use of cost curves to evaluate their potential for implementation of climate change mitigation. Ecological Economics 69, 469-477.

van Kooten, G.C., Eagle, A.J., Manley, J., Smolak, T., 2004. How costly are carbon offsets? A meta-analysis of carbon forest sinks. Environmental Science and Policy 7, 239-251.

van Kooten, G.C., Shaikh, S.L., Suchanek, P., 2002. Mitigating climate change by planting trees: the transaction costs trap. Land Economics 78, 559-572.

van Minnen, J.G., Strengers, B.J., Eickhout, B., Swart, R.J., Leemans, R., 2008. Quantifying the effectiveness of climate change mitigation through forest plantations and carbon sequestration with an integrated land-use model. Carbon Balance and Management 3, 1-20.

Williams, K., 2011. Relative acceptance of traditional and non-traditional rural land uses: views of residents in two regions, southern Australia. Landscape and Urban Planning 103, 55-63.

Winsten, J., Walker, S., Brown, S., Grimland, S., 2011. Estimating carbon supply curves from afforestation ofagricultural land in the Northeastern US. Mitigation and Adaptation Strategies for Global Change 16 (8) 925-942.

Yemshanov, D., Biggs, J., McKenney, D.W., Lempriere, T., 2012. Effects of permanence requirements on afforestation choices for carbon sequestration for Ontario, Canada. Forest Policy and Economics 14 (1) 6-18.

Zomer, R.J., Trabucco, A., Bossio, D.A., Verchot, L.V., 2008. Climate change mitigation: a spatial analysis of global land suitability for clean development mechanism afforestation and reforestation. Agricultural, Ecosystems and Environment 126, 67-80.