Simulation of decorrelated factors in presence of secondary data
Geostatistical simulation of variables with complex multivariate relationships in presence of exhaustive secondary variables such as remotely sensed geophysical measurements is addressed. The primary variables being simulated may be transformed to independent Gaussian factors to partially manage the...
Saved in:
Published in | Spatial statistics Vol. 33; p. 100385 |
---|---|
Main Authors | , , |
Format | Journal Article |
Language | English |
Published |
Elsevier B.V
01.10.2019
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | Geostatistical simulation of variables with complex multivariate relationships in presence of exhaustive secondary variables such as remotely sensed geophysical measurements is addressed. The primary variables being simulated may be transformed to independent Gaussian factors to partially manage the complex relationships, yet they cannot be simulated independently. Although independent, the transformed factors remain dependent on the secondary data and must be simulated in a dependent manner. This interesting aspect of multivariate geostatistics is explored and a hierarchical algorithm for cosimulation using variable amalgamation and intrinsic collocated cokriging is developed. A number of examples are shown to illustrate the multivariate structure of primary data in presence of secondary data.
•Decorrelation of variables in presence of exhaustive secondary data is addressed.•Conditional correlation is used to reproduce the correlations among all variables.•An example with PPMT and ICCK demonstrates the concepts. |
---|---|
ISSN: | 2211-6753 2211-6753 |
DOI: | 10.1016/j.spasta.2019.100385 |