Geoscience Reference
In-Depth Information
was originally developed by von Thünen for urban areas where property owners
seek to optimise their location by trading off access to the urban centre with land
rents. The model is equilibrial and spatially homogeneous, and (perhaps unsurpris-
ingly) fails to reproduce observed patterns of city growth adequately; instead it
produces concentric rings refl ecting the balance between land value and transporta-
tion costs (Bockstael, 1996; Irwin and Geoghegan, 2001; Brown, 2006).
A few recent microeconomic models have departed from some of these restric-
tive assumptions and have adopted a spatially explicit perspective. For example,
Bockstael (1996) and Irwin and Geoghegan (2001) describe a spatially explicit
model of the economics of land-use conversion in the Patuxent watershed in north-
east Maryland, USA. This region is a heterogeneous mix of rural and urban land
uses and is undergoing rapid urbanisation, precisely the type of situation that con-
founds non-spatial bid-rent models. In Bockstael's model, land owners make deci-
sions about whether or not to change land use at a given site on the basis of the
future stream of returns to the parcel given how it is currently used (taking into
account conversion costs). Because knowledge surrounding these decisions is imper-
fect, this decision-making process is framed as discrete probability choices. If there
are n categories of land use, then there are n 2 decisions that land owners could
potentially make. Bockstael (1996) reduces this to just one choice: whether or not
to convert a land parcel from being undeveloped to developed. Thus, the model
requires two pieces of empirical information: (i) the value of each parcel of land
under any possible uses and (ii) the probability of conversion given those land values
and associated conversion costs. To estimate these, Bockstael used an empirical
model of land values (what economists term a 'hedonic pricing model') in which
spatial factors such as neighbourhood conditions were included as drivers of land
value, alongside more usual economic determinants of land value such as parcel size
and access to transport infrastructure. The outcome of this model is a static map
of probabilities of change. Using this framework, the implications of different public
policy scenarios can be explored, as they infl uence the hedonic model, and the
resultant probability maps compared. Subsequent extensions to the model (see Irwin
and Geoghegan, 2001) made it temporally dynamic by incorporating a term that
describes the optimal timing of the decision to convert land.
Although analytical approaches grounded in microeconomic theory have proven
useful, they represent a different direction to that taken by geography and other
disciplines (Drechsler et al., 2007). One of the key criticisms of such microeconomic
models is the assumption that those involved in represent Homo economicus - the
perfectly rational and informed decision maker. Furthermore, the emphasis in
econometrics has largely been on temporal change and on equilibrial conditions
(although spatial econometric tools are being developed - Irwin and Geoghegan,
2001). Again, these research directions are somewhat different to those taken in
geography where the emphasis on space and disequilibrial conditions makes the use
of analytical models problematic.
Empirical-statistical models
Empirical-statistical models, and in particular, a multitude of regression-derived
approaches, have been widely applied for modelling LUCC. These regression
approaches have been criticised on heuristic and methodological grounds; Brown
et al. (2004, p. 401) identify some general problems with empirical-statistical
Search WWH ::




Custom Search