Models that simulate land-use patterns often use either inductive, data-driven approaches or deductive, theory-based methods to describe the relative strength of the social, economic and biophysical forces that drive the various sectors in the land system. An integrated framework is proposed here that incorporates both approaches based on a unified assessment for local land suitability following a monetary, utility-based logic. The framework is illustrated with a hedonic pricing analysis of urban land values and a net present value assessment for agricultural production system in combination with statistics-based assessments of land suitability for other sectors. The results show that limited difference exists between the most commonly applied inductive approaches that use either multinomial or binomial logistic regression specifications of suitability. Land-use simulations following the binomial regression based suitability values that were rescaled to bid prices (reflecting relative competitiveness) perform better for all individual land-use types. Performance improves even further when a land value based description of urban bid prices is added to this approach. Interestingly enough the better fitting description of suitability for urban areas also improves the ability of the model to simulate correct locations for business estates and greenhouses. The simulation alternatives that consider the net present values for agricultural types of land use show the relevance of this approach for understanding the spatial distribution of these types of land use. The combined use of urban land values and net present values for agricultural land use in defining land suitability performs best in our validation exercise. The proposed methodology can also be used to incorporate information from other research frameworks that describe the utility of land for different types of use.
|Journal||Computers, Environment and Urban Systems|
|Publication status||Published - 2015|