| |
| Spatial Statistics |
 |
|
|
| Spatial statistics is the application of statistical theory and techniques to the description and modelling of spatial referenced data. Examples of spatial data are soil or groundwater contamination locations, climate station data (temperature, rainfall, etc…), crop yields, housing prices, spatial incidence of certain types of diseases or any data that is referenced to some x-y coordinate system.
Spatial statistics is very much an applied subject within the field of statistics; its development has been the work of mining engineers, soil scientists, geologists as well as statisticians. There are applications in epidemiology, forestry, atmospheric sciences, hydrology, geophysics, global change research, geography and the social sciences.
Geostatistics is a subset of spatial statistics. In geostatistics, the observations in some x-y coordinate system are potentially available at any location within the region of interest, but the data are obtained only for a sample of such locations. Engineers and environmental scientist routinely encounter estimation problems when analyzing data from field observations.
For example, measurements of mercury concentration are sampled from soil locations in a region of western Quebec. The tools of geostatistics can help in its spatial estimation. The semivariogram provides description of the spatial dependence and spatial structure of the observations. Kriging may be used to predict the mercury concentration at the unsampled sites, from which a contour map of predicted values is obtained, summarizing the pattern of variation of mercury. Also assessments may be made of the spatial uncertainty of the results. Other examples are the spatial description of hazardous waste sites, risk assessment of ground water contamination and atmospheric ozone distribution. |
|
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
Multivariate Data Analysis School |
|
|
|
|
|
|
|
| |
