|Spatial Interpolation of Environmental Parameters with Copulas|
|Project manager:||Prof. Dr. rer.nat. Dr.-Ing. András Bárdossy, Dr.-Ing. Claus Haslauer, M.Sc.|
|Research assistants:||Dr.-Ing. Claus Haslauer, M.Sc.|
|Duration:||1.5.2007 - |
|Funding:||German Research Foundation|
Abstract:The proposed research involves developing comprehensive methods for using Copulas to spatially interpolate environmental parameters. Problems with traditional geostatistical methods have been recognized. Copulas are expected to solve some of these drawbacks, particularly drawbacks related to uncertainty estimates.
A Copula is an effective way to describe the dependence of multivariate distributions with any kind of marginal distributions. Copulas can be regarded as the pure expression of dependence without the influence of a marginal distribution. Furthermore, Copulas enable the calculation of the conditional marginal distribution corresponding to the point for which the estimation is carried out. The conditional distribution can be used to estimate confidence intervals for the estimator. These confidence intervals are based on the dependence structure described by the Copula, the spatial configuration of the observation points, and their observed values. When using Copulas, the same estimated value can occur from the same spatial configuration with different confidence intervals. These features are a substantial enhancement compared to those offered by traditional geostatistical approaches such as Kriging.