Issues in spatially explicit statistical land-use/cover change (LUCC) models: Examples from 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)
Under the hood: issues in the specification and interpretation of spatial regression models
Agricultural Economics
(2002)Modelling land use change with generalized linear models—a multi-model analysis of change between 1860 and 2000 in Gallatin Valley, Montana
Journal of Environmental Management
(2004)- et al.
Modeling the relationships between land use and land cover on private lands in the Upper Midwest
Journal of Environmental Management
(2000) - et al.
Extensification of agriculture and deforestation: empirical evidence from Sudan
Agricultural Economics
(1994) - et al.
Multi-scale analysis of a household level agent-based model of landcover change
Journal of Environmental Management
(2004) - et al.
Household determinants of deforestation by Amerindians in Honduras
World Development
(1997) Multilevel modeling on farmland distribution in Japan
Land Use Policy
(2001)- et al.
Land use externalities, growth management policies, and urban sprawl
Regional Science and Urban Economics
(2004) - et al.
Land use dynamics in the Central Highlands of Vietnam: a spatial model combining village survey data and satellite imagery interpretation
Agricultural Economics
(2002) - et al.
The dynamics of land-cover change in western Honduras: exploring spatial and temporal complexity
Agricultural Economics
(2002)
Deforestation and land use change: sparse data environments
Agricultural Economics
Adaptation to climate variability and change in the US Great Plains: a multi-scale analysis of Ricardian climate sensitivities
Agriculture, Ecosystems and Environment
Land-cover change model validation by an ROC method for the Ipswich watershed, Massachusetts, USA
Agriculture, Ecosystems and Environment
Detecting important categorical land changes while accounting for persistence
Agriculture, Ecosystems and Environment
Proximate causes of land-use change in Narok District, Kenya: a spatial statistical model
Agriculture, Ecosystems and Environment
The causes of tropical deforestation in Ecuador: a statistical analysis
World Development
Land cover change and landscape fragmentation—comparing the utility of continuous and discrete analyses for a western Honduras region
Agriculture, Ecosystems and Environment
Predicting land-use change
Agriculture, Ecosystems and Environment
A multiscale analysis of land use land cover change and NDVI variation in Nang Rong district, Northeast Thailand
Agriculture, Ecosystems and Environment
Spatial scale problems and geostatistical solutions
Professional Geographer
Modeling economics and ecology: the importance of a spatial perspective
American Journal of Agricultural Economics
Land-use patterns and water quality: the effect of differential land management controls
Roads, lands use, and deforestation: a spatial model applied to Belize
World Bank Economic Review
Cited by (84)
The effects of climate and forest cover variability on the hydrological regulation of an eastern Andean Cusiana river sub-basin
2023, Ecohydrology and HydrobiologyUrbanization, economic development, environmental and social changes in transitional economies: Vietnam after Doimoi
2019, Landscape and Urban PlanningConstructing the ecological security pattern of the Daqing River Basin for the high-standard ecological construction of Xiong'an New Area, China
2024, Environment, Development and SustainabilityHistory, neighborhood, and proximity as factors of land-use change: A dynamic spatial regression model
2024, Environment and Planning B: Urban Analytics and City Science