Available with Geostatistical Analyst license.
The general context in which the Geostatistical Analyst tools are used is that there is some phenomenon occurring in the landscape that is of interest (for example, contamination of soil, water, or air by a pollutant; the content of gold or some other metal in an area that you wish to mine; and so on). Exhaustive studies are expensive and time-consuming, so the phenomenon is usually characterized by taking samples at different locations. Geostatistics is then used to produce predictions for the unsampled locations. Geostatistical Analyst offers tools to help explore the sample data, as well as several different interpolation methods.
Interpolation methods produce only one value for every location in the study area. This value may not provide all the information needed to make good decisions. Additionally, the interpolated surfaces are usually smoother (less variable) than the real phenomenon.
In contrast, simulation techniques generate many interpolated surfaces, all of which replicate the spatial characteristics found in the sample data. When all the simulated surfaces are assembled, they provide a distribution of values for each location in the study area. These distributions are the basis for risk analyses, economic decision making, and other estimations involving uncertainty and allow analysts to make more informed decisions.