hdeco (version 0.4.1)

CARsimu: Conditional AutoRegressive surface simulator

Description

A stationary tool for simulating a surface (as a matrix of values) with specified spatial autocorrelation parameters. The simulation return $2^level*2^level$ real-valued surface (e.g., landscape) based on the FFT algorithm and the spectral (or autocorrelation theorem using first and second order neighbours in N-S, E-W, NW-SE, and NE-SW directions.

Usage

CARsimu(level = 5, row1 = 0.2499999, row2 = 0, col1 = 0.2499999, col2 = 0, rc1 = 0, cr1 = 0)

Arguments

level

Integer: controls the dimensions of the output surface ($2^level*2^level$)

row1

First order neighbour East-West autocorrleation parameter

row2

Second order neighbour East-West autocorrelation parameter

col1

First order neighbour North-South autocorrelation parameter

col2

Second order neighbour North-South autocorrelation parameter

rc1

First order neighbour NW-SE autocorrelation parameter

cr1

First order neighbour NE-SW autocorrelation parameter

Value

• Returns a $2^level*2^level$ matrix of real numbers. It is possible to use the image() or imaks() functions to graphically display the surface. Saving the output from this function into a new object is likely the most desireable usage. Ths surface can then be factored to produce binary maps with specified proportions as in Remmel and Csillag (2003).

Details

The sum of all six autocorrelation parameters MUST be less than 0.5. Their effect is highly non-linear, thus there is a marked difference beetween 0.499 and 0.499999. This implementation corresponds to the homogeneous (stationary) conditional autoregressive (CAR) model.

References

Remmel, T.K. and F. Csillag. 2003. When are two landscape pattern indices significantly different? Journal of Geographical Systems 5(4):331-351.

Examples

carimage <- CARsimu(level=5, row1=0.2499999, row2=0, col1=0.2499999, col2=0, rc1=0, cr1=0)
image(carimage)