The script generates spatial data with a scale-invariant power spectrum (1/f noise) and a normal error distribution.
The spectral density of the data is proportional to f^BETA, where f is the frequency and BETA is the spectral exponent (BETA=0 is white noise, BETA=-2 is Brown noise).

To generate Brown noise on a 10x10 grid you type

spatialPattern([10,10],-2)

which gives a 10x10 matrix containihng the data.

1/f noise is not stationary, and so doesn't have a well defined variance. A variogram of the data show variance linearly increasing with increasing lag.

Time series data can also be generated. For example, a pink noise time series of lenth 1000 is produced by

spatialPattern([1000,1],-1)