Efficient Approximation of the Spatial Covariance Function for Large Datasets - Analysis of Atmospheric CO2 Concentrations
Linear mixed effects models have been widely used in the spatial analysis of environmental processes. However, parameter estimation and spatial predictions involve the inversion and determinant of the n times n dimensional spatial covariance matrix of the data process, with n being the number of observations. Nowadays environmental variables are typically obtained through remote sensing and contain observations of the order of tens or hundreds of thousand on a single day, which quickly leads to bottlenecks in terms of computation speed and requirements in working memory. Therefore techniques for reducing the dimension of the problem are required. The present work analyzes approaches to approximate the spatial covariance function in a real dataset of remotely sensed carbon dioxide concentrations, obtained from the Atmospheric Infrared Sounder of NASA's 'Aqua' satellite on the 1st of May 2009. In a cross-validation case study it is shown how fixed rank kriging, stationary covariance tapering and the full-scale approximation are able to notably speed up calculations. However the loss in predictive performance caused by the approximation strongly differs. The best results were obtained for the full-scale approximation, which was able to overcome the individual weaknesses of the fixed rank kriging and the covariance tapering.
|Date of creation:||Aug 2013|
|Contact details of provider:|| Postal: Grosse Scharrnstrasse 59, 15230 Frankfurt (Oder)|
Phone: +49 (0)335 5534 2387
Fax: +49 (0)335 5534 2516
Web page: http://www.europa-uni.de/recap15/
More information through EDIRC
References listed on IDEAS
Please report citation or reference errors to , or , if you are the registered author of the cited work, log in to your RePEc Author Service profile, click on "citations" and make appropriate adjustments.:
- Noel Cressie & Gardar Johannesson, 2008. "Fixed rank kriging for very large spatial data sets," Journal of the Royal Statistical Society Series B, Royal Statistical Society, vol. 70(1), pages 209-226.
- Kaufman, Cari G. & Schervish, Mark J. & Nychka, Douglas W., 2008. "Covariance Tapering for Likelihood-Based Estimation in Large Spatial Data Sets," Journal of the American Statistical Association, American Statistical Association, vol. 103(484), pages 1545-1555.
- Huiyan Sang & Jianhua Z. Huang, 2012. "A full scale approximation of covariance functions for large spatial data sets," Journal of the Royal Statistical Society Series B, Royal Statistical Society, vol. 74(1), pages 111-132, 01.
When requesting a correction, please mention this item's handle: RePEc:euv:dpaper:009. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Daniel Becker)
If references are entirely missing, you can add them using this form.