The physical, economic and policy drivers of land conversion to forestry in Ireland

https://doi.org/10.1016/j.jenvman.2013.10.017Get rights and content

Highlights

  • Forest cover has been expanding in Ireland due to state policies but targets have not been met.

  • Annual afforestation per electoral division was modelled using panel and spatial panel models.

  • Physical, economic and policy related variables were found to influence afforestation.

  • Soil quality was noted as a particularly important variable but this effect has changed over time.

  • Expansionary agricultural and restrictive environmental policies may act as barriers to afforestation.

Abstract

Land use change is fundamentally a product of the interaction of physical land characteristics, economic considerations and agricultural and environmental policies. Researchers are increasingly combining physical and socio-economic spatial data to investigate the drivers of land-use change in relation to policy and economic developments. Focusing on Ireland, this study develops a panel data set of annual afforestation over 2811 small-area boundaries between 1993 and 2007 from vector and raster data sources. Soil type and other physical characteristics are combined with the net returns of converting agricultural land to forestry, based on the micro-simulation of individual farm incomes, to investigate land conversion. A spatial econometric approach is adopted to model the data and a range of physical, economic and policy factors are identified as having a significant effect on afforestation rates. In addition to the financial returns, the availability and quality of land and the implementation of environmental protection policies are identified as important factors in land conversion. The implications of these factors for the goal of forest expansion are discussed in relation to conflicting current and future land use policies.

Introduction

Land-use change modelling requires combining both physical and economic spatial data if it is to be used to understand policy developments and predict future land-use changes (Seto and Kaufmann, 2003). In the absence of data concerning the economic implications of land-use decisions, interpreting historic change, particularly in relation to policy developments, poses a significant challenge (Bockstael, 1996). Although physical drivers of land conversion may be identified, the causal relationship between characteristics and change may be less clear (Irwin and Geoghegan, 2001). This is perhaps of most relevance in enterprises where state and regional policies have a defining and widespread impact, such as agriculture and forestry. Despite the recognition of the importance of including economic data in spatial models researchers may be constrained by the existence of data or the scale at which data are available. In agricultural research, spatial data on farm incomes at the individual or local level may be limited. One approach to overcoming this issue is to simulate individual farm data from broader regional or national data (O'Donoghue et al., 2012).

Increasing forest cover is a common goal internationally and has been supported within European agricultural policy for a number of decades (Nijnik and Bizikova, 2008). Land conversion to forestry is a complex issue that is influenced by social, economic and environmental factors that policy-makers should account for in the development of forest policy and the setting of targets (Beach et al., 2005). Thus, understanding afforestation requires combining multiple sources of data within a modelling approach that ideally accounts for both the spatial and temporal nature of the phenomenon. Spatial econometric models offer the potential to investigate and quantify the effects of these factors on land conversion while explicitly addressing the spatial nature of the data (Radeloff et al., 2012).

Afforestation is increasingly valued for its potential to enhance ecosystem services and is being actively promoted in many countries through state policy and support (Kanowski, 2010). Forest cover expansion is included as a source of carbon dioxide emission reduction under the Kyoto Protocol, which is a significant factor in the promotion of forest expansion policies (Nijnik and Bizikova, 2008). Similar to many countries, Ireland has sought to increase its forest cover for some time with rural employment and economic diversification benefits being important drivers in the 20th century and ecosystem services being increasingly recognised in modern forest policy (Department of Agriculture, Food and Forestry, 1996, OCarroll, 2004).

Ireland offers a particularly interesting example of forest expansion policy as it possesses one of the lowest areas of forest cover in Europe, despite possessing excellent growing conditions for commercial forestry, and a history of ambitious afforestation policies (OCarroll, 2004). Current forest cover stands at 10.9% with the majority of this area composed of plantation forests established in the last hundred years. The goal of state policy is to increase forest cover to 17% by the year 2030 through private planting (Department of Agriculture, Food and Forestry, 1996). Historical afforestation policies and establishment in Ireland have a distinctive locational bias defined by the quality of the underlying land (Upton et al., 2012). Initial efforts by the state to expand forest cover were enthusiastic but poorly planned and resulted in relatively low levels of planting (OCarroll, 2004). Planting was limited to sub-marginal land, often at higher elevations with peat soils. Although grants for planting by private landowners were available, private afforestation was limited until the late 1980s when annual premiums were introduced under the Western Package Scheme which was co-funded by the EU (EU Regulation No. 1820/80). These payments compensated private landowners, for a limited period of time, for lost agricultural income as forests developed. This resulted in a significant increase in afforestation by private landowners (Fig. 1). Supports for planting by state agencies were removed in the mid-1990s, which essentially saw the end of public planting. Initially policies for private planting specifically targeted agriculturally disadvantaged parts of Ireland. Since 1992 a consistent policy of grants and annual premiums for 20 years open to all private landowners, but with higher rates for farmers, has been in place. Ireland benefited from funding for afforestation by the EU under the Community aid scheme for afforestation from 1992 (Council Regulation (EEC) No 2080/92) and under support for rural development from 2000 (Council Regulation (EC) No 1257/1999). The availability of grants and premiums makes forestry a financially attractive enterprise for many farmers but particularly those engaged in extensive livestock rearing (Breen et al., 2010). However, annual afforestation rates have been variable and declining since 2005.

Plantation forests can achieve high productivity rates even on poorly drained mineral soils (Farrelly et al., 2011), giving forestry a greater competitive advantage on poorer quality soils. Nonetheless, farmers have been reluctant to plant forestry due to a range of factors, including the non-pecuniary costs, related to a change in land use and lifestyle. Although the Irish public support and value afforestation greatly, farmers may view forestry as a less desirable land use (Upton et al., 2012). Land conversion to forest by private landowners is a complex issue with multiple underlying causes, including, but not limited to, the incentives and restrictions of state policies (Beach et al., 2005). The effects of policy changes and market conditions on afforestation rates in Ireland have been explored using time-series and panel data (McKillop and Kula, 1987, McCarthy et al., 2003). In general such studies find that the profitability of agriculture and forestry are significant factors in determining afforestation rates. Researchers have examined afforestation in Ireland on the county level but failed to account for the spatial nature of the data in the modelling process or the physical characteristics of the land (McCarthy et al., 2003). Examinations of private afforestation in Ireland have shown that land quality is a defining aspect of the decision-making process by farmers (Ní Dhubháin and Gardiner, 1994, Howley et al., 2012). Land quality underlies the productivity and profitability of alternative land uses, making it an essential element in understanding land conversion. In addition, forestry has been recognised as an enterprise only “suitable” for the worst quality land by landowners (O'Leary et al., 2000). This may be driven by the belief that land should be used for the production of food if at all possible rather than an aversion to forestry per se (McDonagh et al., 2010). However, strong negative views of afforestation have been identified in parts of Ireland, particularly those that saw a rapid expansion of forest cover over a relatively short time-period (O'Leary et al., 2000).

It has been suggested that conservation policies related to protected habitats or species have reduced annual afforestation rates and discouraged applications from relevant areas (Collier et al., 2002). The EU habitats (92/43/EEC) and birds (79/409/EEC) directives resulted in the identification of special areas of conservation and special protection areas, which complemented the Irish specification of natural heritage areas. Habitats and species related to these areas are given legal protection and applications for afforestation funding within these areas require approval from the Irish National Parks and Wildlife Service. Forests can increase soil acidity through their capacity of trees to scavenge industrial air pollutants or sea-salts (Dunford et al., 2012). Where this occurs on soils with poor buffering capacity adjacent water-ways may become acidified. The Forest Service in Ireland has identified areas that are considered at risk of acidification due to the poor buffering capacity of the soil and afforestation is controlled in these areas.

Spatial models of land-use change are employed to gain greater insight into the drivers of change, the effectiveness of policies and to predict future land conversion (Lubowski et al., 2008). Land-use change studies have been conducted on a diverse range of issues including urban expansion (Seto and Kaufmann, 2003), deforestation (Wyman and Stein, 2010) and afforestation (Clement et al., 2009). Land quality, related to factors such as soil, elevation and slope, is one of the essential determinants of private land-use decision-making given its underlying effect on productivity and should be incorporated into spatial models (Lubowski et al., 2008). Soil type and other physical characteristics have been identified as significant factors in land-use change models (Fu et al., 2006, Chakir and Parent, 2009). Ultimately, however, the financial implications of land-use change should be included in models if the decisions made by private landowners are to be understood within an economic framework (Bockstael, 1996).

In developing spatial models of land-use change, researchers generally employ satellite imagery from different time-periods and explore change at the single land-parcel or pixel level over a set period (e.g. Radeloff et al., 2012). Alternatively, researchers may examine total changes across administrative boundaries which can facilitate the incorporation of economic data more readily (Seto and Kaufmann, 2003). In modelling spatially derived data researchers should test for spatial autocorrelation amongst the observations, which can lead to biased estimations (Anselin, 2010). Spatial dependence amongst the observations is considered one of the primary problems with employing spatially explicit panel data and a number of approaches to dealing with this potential source of bias have been developed (Elhorst, 2003). One approach is to specify a spatial lag variable that accounts for the interaction of the dependent variable in related observations. This requires the specification of the spatial relationship between observed units, which can be expressed in a spatial weights matrix.

Understanding the drivers of afforestation should assist in explaining afforestation patterns and help to inform meaningful forest policy. Afforestation by private landowners may be affected by a combination of market drivers, policy variables, owner characteristics and land conditions (Beach et al., 2005). In the context of this study, it is hypothesised that the underlying characteristics of the land, the financial implications of conversion and the constraining effects of conservation policies influence afforestation. Thus, the primary aim of the study is to test the nature of these effects in explaining afforestation in Ireland and their significance to forest and broader land-use policies. Geographic information system (GIS) analysis, the micro-simulation of farm-level incomes and financial analysis techniques are employed to build a panel dataset to explore the importance of physical, economic and policy related factors in explaining annual afforestation in Ireland between 1993 and 2007. A random effects and a spatial autoregressive random effects model that accounts for the spatial correlation of observations are employed to model the data.

Section snippets

Methodology

The boundaries of electoral divisions (EDs) were employed as the spatial unit in which observations would be specified as they represent the smallest spatial unit for which economic data are available. Ireland is divided into 3440 EDs in total but those which occur within cities and those for which agricultural data were not available were removed, resulting in a sample of 2811 (Fig. 2(a)). Employing a GIS these boundaries were intersected with available spatial data, including grant-aided

Results

Correlation between the independent variables is generally low except between the soil variables, which is expected given that they are proportional to each other (Table 4). Thus, multi-collinearity was not deemed to be a significant issue in the models.

All included variables had a significant effect on afforestation and there are no major changes in the sign or scale of coefficients between models (Table 5). However, the coefficient of the spatial lag is significant and positive indicating

Discussion

The results highlight the importance of underlying physical land characteristics in understanding afforestation. Physical site characteristics, such as soil and elevation, are essential factors in understanding the natural distribution of forests (Felicísimo et al., 2002) and have been shown to be important predictors of land-use change such as land abandonment (Sluiter and de Jong, 2007) and forest expansion (Fu et al., 2006). Such findings highlight the limitations imposed by site quality on

Conclusion

Afforestation by private landowners is generally seen as a function of agricultural commodity and timber prices, land prices and government subsidisation. However, fundamental of understanding this land-use change is the influence of physical characteristics of the land, particularly soil quality. Commercial forestry is less reliant on site quality than other potential land uses and high productivity levels can be attained in areas considered marginal for agriculture. This study demonstrates

Acknowledgements

Funding for this research was provided by COFORD, Department of Agriculture, Food and the Marine.

References (45)

  • V. Upton et al.

    Preferences and values for afforestation: The effects of location and respondent understanding on forest attributes in a labelled choice experiment

    For. Policy Econ.

    (2012)
  • M.S. Wyman et al.

    Modeling social and land-use/land-cover change data to assess drivers of smallholder deforestation in Belize

    Appl. Geogr.

    (2010)
  • L. Anselin

    Thirty years of spatial econometrics

    Pap. Reg. Sci

    (2010)
  • K. Black et al.

    The extent of recent peatland afforestation in Ireland

    Ir. For.

    (2008)
  • N.E. Bockstael

    Modeling economics and ecology: the importance of a spatial perspective

    Am. J. Agric. Econ.

    (1996)
  • J. Breen et al.

    Irish land use change and the decision to afforest: an economic analysis

    Ir. For.

    (2010)
  • E. Brockerhoff et al.

    Plantation forests and biodiversity: oxymoron or opportunity?

    Biodivers. Conserv.

    (2008)
  • E. Buscardo et al.

    The early effects of afforestation on biodiversity of grasslands in Ireland

    Biodivers. Conserv.

    (2008)
  • M.S. Carroll et al.

    Back where they once belonged? Local response to afforestation in County Kerry, Ireland

    Sociol. Rural.

    (2011)
  • R. Chakir et al.

    Determinants of land use changes: a spatial multinomial probit approach

    Pap. Reg. Sci.

    (2009)
  • J.P. Clinch

    Economics of Irish Forestry: Evaluating the Returns to Economy and Society

    (1999)
  • P. Collier et al.

    Factors Influencing Farmer Participation in Forestry

    (2002)
  • Cited by (29)

    • Trends and influential factors of high ecological status mobility in Irish Rivers

      2022, Science of the Total Environment
      Citation Excerpt :

      The FIPS-IFS dataset was based on field sampling and was created using GIS and remote sensing, with a mapping unit of 1 ha. The variable used was annual afforestation, reflecting the percentage area of newly planted forest in each ED (Upton et al., 2014). Apart from human activities, the physical environment, including soil type, bedrock geology, landscape, hydrological and climatic factors all affect river water quality (Silva and Williams, 2001; Moses et al., 2011; Rankinen et al., 2015).

    • Spatial dependence and determinants of conservation easement adoptions in the United States

      2021, Journal of Environmental Management
      Citation Excerpt :

      Spatial econometric models can assess the degree of spatial dependence in a dataset and incorporate it into the modeling through spatially lagged terms. It has been applied to several issues related to natural resources and the environment with spatial panel data, such as tree planting (Li and Zhang, 2007), land conversion decisions (Upton et al., 2014), and urban pollution (Cheng et al., 2020). In the current application, the first step will be assessing whether there is spatial autocorrelation (dependence) in the distribution of conservation easements across the United States.

    • Addressing the challenge of wood mobilisation through a systemic innovation lens: The Irish forest sector innovation system

      2021, Forest Policy and Economics
      Citation Excerpt :

      These forecasts assume that private forests, planted since the mid-eighties with the support of afforestation capital grants and afforestation subsidies (DAFM, 2020), will be managed commercially to maximise net present value by employing the same intensive thinning/clearfell silvicultural system as state managed (Coillte) forests (Ní Dhubháin et al., 2015; Phillips et al., 2016). Unlike most other European States, private forest ownership in Ireland is dominated by new forest owners, the majority of who are farmers (81%) engaging in forestry for the first time (Duesberg et al., 2014; Upton et al., 2014; DAFM, 2020). These new owners have little knowledge or experience of farm forest management, in contrast to forest owners in countries with a longer tradition of farm forestry (Weiss et al., 2018) and considerable uncertainty surrounds both their intentions and capacity to actively manage their forests (Upton et al., 2017).

    View all citing articles on Scopus
    View full text