Towards hyper-dimensional variography using product-sum covariance model
Authors
Jovan Tadic — Lawrence Berkeley National Laboratory
Sebastien Christophe Biraud — Lawrence Berkeley National Laboratory
Category
General topics
Description
Average kriged temperature for Oklahoma (Jan 2003) using hyper-dimensional kernel
We present a new method for modeling hyper-dimensional covariance (variogram) structures using the product-sum covariance model initially developed to characterize spatio-temporal variability. This method can be used to model variability in anisotropic conditions with multiple axes of anisotropy or when temporal evolution is involved, and thus is applicable to “full anisotropic 3D+time” conditions (for instance latitude, longitude, altitude, and time). We applied this method to generate distributed surface meteorological fields (temperature is shown as an example) between 2000 and 2015, using Oklahoma Mesonet sites and ARM Southern Great Plains observations. These distributed meteorological fields could augment the ARM best-estimate datastream (ARMBE) and contribute to the development of routine Large-Eddy Simulation (LES) modeling framework over the ARM SGP.
