Elsevier

Land Use Policy

Volume 24, Issue 3, July 2007, Pages 521-530
Land Use Policy

Issues in spatially explicit statistical land-use/cover change (LUCC) models: Examples from western Honduras and the Central Highlands of Vietnam

https://doi.org/10.1016/j.landusepol.2005.09.007Get rights and content

Abstract

Land-use/cover change (LUCC) results from the complex interaction of social, ecological and geophysical processes. Land users make decisions about their environment that are governed and influenced by political and institutional constraints at local, regional, national and international levels. Statistical analysis of LUCC phenomena is one powerful tool due to its ability to test theoretical assumptions, rank relative factors, and yield rigorous hypotheses test. However, due to the complex nature of coupled human–environment systems, LUCC statistical modeling presents conceptual as well as technical challenges. Careful consideration of these challenges, as well as implementing approaches to deal with them, is necessary in order for such models to inform policy and practice. As examples, we present illustrations of two statistical models of land use in the mountains of western Honduras and the Central Highlands of Vietnam.

Introduction

Land-use/cover change (LUCC) results from the complex interaction of social, ecological and geophysical processes. Land users make decisions about their environment that are governed and influenced by political and institutional constraints at local, regional, national and international levels. Despite much work studying LUCC and LUCC processes, a definitive understanding of LUCC is elusive, and there is no universal set of guidelines for policy that could mitigate LUCC across the board; the local situation and context are critical (Geist and Lambin, 2002).

LUCC researchers are increasingly realizing that an integrated set of techniques is required to understand these phenomena. In addition to process-based modeling and the use of case studies in examining LUCC (Veldkamp and Lambin, 2001), statistical analysis is a powerful tool due to its ability to test theoretical assumptions, rank relative factors, and yield rigorous hypotheses tests. Statistical analyses of LUCC present the opportunity to link the general and the specific, in that one can effectively identify the relative contribution of broader factors such as institutional and market forces, while controlling for factors particular to the location. However, modeling LUCC processes requires integrating data across space, time and level of analysis, and there remain barriers to broader implementation of best-practice techniques (Rindfuss et al., 2004).

The underlying objective of this paper is to examine statistical tools and techniques most useful to studying LUCC, and to assess current best practice. The structure of this paper is as follows. We begin with examining conceptual challenges researchers face in the implementation of LUCC statistical models and the myriad technical difficulties that arise as a result (Rindfuss et al., 2004). We then explore the means available to overcome these problems, with reference to key LUCC statistical models from the current literature. Finally, we present two empirical examples, from study areas in western Honduras and the Central Highlands of Vietnam that make use of some of the issues. We end with a final discussion of the main impediments and challenges to broader implementation of best practice techniques.

Section snippets

Conceptual challenges

This section examines the tradeoffs that are implicit in the empirical representation of LUCC processes and the implications for statistical analysis. LUCC is necessarily conceptualized as comprising a dynamic, coupled human–natural system. Land-use change, such as urbanization or deforestation, can have profound influences on ecosystems, but ultimately, ecosystem responses influence further land-use change. For example, areas proximate to a lake may be prime real estate, spurring development,

Empirical challenges

In addition to conceptual challenges, LUCC statistical models generally suffer from a variety of integration and specification problems. Initially, multiple regression was the primary means to link hypothesized driving forces of land use to observed land-cover patterns. More recently, researchers have attempted to employ statistical techniques as complementary tools to integrated, mixed method approaches (Brown et al., 2000; Dutcher et al., 2004; Parker et al., 2003). Multiple regression is

Analysis

In the next sections, we demonstrate techniques that could potentially extend the inferential capabilities of a multinomial logit using two empirical examples. The first uses repeated spatial sampling in order to examine whether estimated areas at risk are robust. The second explores what additional information can be gleaned by examining estimated probabilities in greater detail. The analysis builds on prior work on Western Honduras (Munroe et al., 2002, Munroe et al., 2004; Southworth et al.,

Discussion

The analysis presented here indicates that statistical models, while representing powerful tools for exploring patterns between land-use/cover processes and their hypothesized covariates, are subject to challenges and pitfalls in their use. Particularly if such models are used to develop policy prescriptions and priorities, it is important to be able to assess qualitatively how these pitfalls may lead to misleading inferences about the nature of these local processes. Serneels and Lambin (2001)

Acknowledgements

We gratefully acknowledge the support of the National Science Foundation (NSF) (SBR-9521918) as part of the ongoing research at the Center for the Study of Institutions, Population, and Environmental Change (CIPEC) at Indiana University; and the Deutsche Forschungsgemeinschaft (DFG) under the Emmy-Noether Program. We would also like to thank Catherine Calder for thoughtful discussion on Bayesian modeling, and Elena Irwin for useful comments. Lastly, we are grateful for the suggestions of two

References (49)

  • G.C. Nelson et al.

    Deforestation and land use change: sparse data environments

    Agricultural Economics

    (2002)
  • C. Polsky et al.

    Adaptation to climate variability and change in the US Great Plains: a multi-scale analysis of Ricardian climate sensitivities

    Agriculture, Ecosystems and Environment

    (2001)
  • R.G. Pontius et al.

    Land-cover change model validation by an ROC method for the Ipswich watershed, Massachusetts, USA

    Agriculture, Ecosystems and Environment

    (2001)
  • R.G. Pontius et al.

    Detecting important categorical land changes while accounting for persistence

    Agriculture, Ecosystems and Environment

    (2004)
  • S. Serneels et al.

    Proximate causes of land-use change in Narok District, Kenya: a spatial statistical model

    Agriculture, Ecosystems and Environment

    (2001)
  • D. Southgate et al.

    The causes of tropical deforestation in Ecuador: a statistical analysis

    World Development

    (1991)
  • J. Southworth et al.

    Land cover change and landscape fragmentation—comparing the utility of continuous and discrete analyses for a western Honduras region

    Agriculture, Ecosystems and Environment

    (2004)
  • A. Veldkamp et al.

    Predicting land-use change

    Agriculture, Ecosystems and Environment

    (2001)
  • S.J. Walsh et al.

    A multiscale analysis of land use land cover change and NDVI variation in Nang Rong district, Northeast Thailand

    Agriculture, Ecosystems and Environment

    (2001)
  • P.M. Atkinson et al.

    Spatial scale problems and geostatistical solutions

    Professional Geographer

    (2000)
  • Bergeron, G., Pender, J.L. 1999. Determinants of land use change: evidence from a community study in Honduras....
  • N.E. Bockstael

    Modeling economics and ecology: the importance of a spatial perspective

    American Journal of Agricultural Economics

    (1996)
  • N. Bockstael et al.

    Land-use patterns and water quality: the effect of differential land management controls

  • K.M. Chomitz et al.

    Roads, lands use, and deforestation: a spatial model applied to Belize

    World Bank Economic Review

    (1996)
  • Cited by (84)

    View all citing articles on Scopus
    View full text