|
Abstract:
|
In environmental science, data are often obtained from computer models or monitoring networks. It is of importance to accommodate the spatial misalignment between the two data sources for calibration of the computer model outputs and for better forecast in the future. In this work, we propose a Bayesian spatial model with spectral methods to capture the relationship between two data sources. The key advantage of our approach is that we can calibrate both the marginal distribution and the spatial correlation of computer model outputs. We apply our methodology to temperature data and show how the model biases can be adjusted.
|
