This is a simulation of Brownian motion in two dimensions. As the iteration progresses, any cross-section through the dataset will resemble a Brownian random walk. The generation process ensures that nearby points will be correlated.

By encoding the height as a color, the dataset is made to resemble a topographic map, although of somewhat higher dimension (more wild) than Earth coastlines. The height is encoded as a byte, and therefore has only 255 values. Therefore, at the extremes of the range (white for highest altitude, deep blue for lowest), clipping occurs which can occasionally be seen.

Java applet by:
Joe Knapp jmk@worldnet.att.net