July 23, 2011

landscape functional zoning methods: cellular automata

Zoning, Landscape zoning or landscape functional zoning is a method, main working mode of spatial planers precisely landscape planners. In short it is a method of theoretically segmenting the land due to it's best using purposes, in other words ascribing the functions to the land like: agricultural land, recreation land, protected land, urbanized land and so on. And it is generally made on national scale by government. But more precisely, many land usage categories are used, due to different methodologies used by different countries simply due to different scope, different criterion for land quality, vegetation, the level of urbanization, that also depends on historical reasons and so on.
Briefly, and maybe more scientifically the concept is described in wikipedia.

Zoning was and still is made using logical methods, cartographically visualizing territory: overlaying many layers (GIS works here perfectly), analyzing and making deductions.
I found one example with zoning schema (picture above) in one of presentations made at EIONET (European Topic Centre on Spatial Information and Analysis). It just shows one particular way, or one approach.

Nowadays, mathematical models are created to help to solve this ambiguate problem.

One of a quantitative methods for zoning protected natural areas are presented in argentinians paper  (Verdiell et all, 2003) that talks about model based on a simulated annealing algorithm.

Another well known computational method that is more and more used in landscape planning, is cellular automata (CA) (check wikipedia), that mostly helps to model the land use change or indeed to simulate the land use change.
Indeed, W. Tobler, the one who defined the first law of geography, was the first proposing to use cellular automata as a tool for modeling spatial dynamics.

But the possible implications of the idea was done by H. Couclelis and her thoughts where written down in the series of theoretical papers. Latter, the approach has been implemented by others in variety of applications, as well as with GIS.
Cellular automata can be thought of as very simple dynamic spatial systems in which the state of each cell in an array depends on the previous state of the cells within a neighborhood of the cell, according to a set of state transition rules.  Because the system is discrete and iterative, and involves interactions only within local regions rather than between all pairs of cells, a CA is very efficient computationally.  It is thus possible to work with grids containing hundreds of thousands of cells.  The very fine spatial resolution that can be attained is an important advantage when modeling land use dynamics, especially for planning and policy applications, since spatial detail represents the actual local features that people experience, and that planners must deal with. (The paper briefly explains the model, it's implication and usage White et al. 2003)
(from: Engelen 2002)
Cellular automata consists of Euclidean space, a cell neighbourhood, a set of discrete cell states, discrete time steps, and transition rules which determine the state of the cell as a function of the states of neighbourhood cells (Engelen 2002White et al. 2003).

There are various papers about modelling land use change integrated with CA. Some I already quoted, the introductory ones. I have noticed that cellular automata is mostly used for modelling urban land change, where a lot of precision is required and celular automata can help for that.
I also found cellular automata is a useful method to follow historical changes. Here is an interesting article: "Exploring the Historical Determinants of Urban Growth Patterns through Cellular Automata" (Stanilov et al. 2011), that analysis West London as an example. The results have shown that the spatial relationships between physical environment and land uses are stable through time, and have small changes due to history behind. And due to that the spatial land development signature can be determined as relationships constitute a basic code for urban growth.

No comments:

Post a Comment