Spatio-temporal prediction inside a free-air CO2 enrichment system

  • Matthew W. Mitchell
  • Marcia L. Gumpertz


The Japanese RiceFACE (Free-Air CO2 Enrichment) project was a three-year investigation into the effect of elevated CO2 on rice. Four rings were built to emit elevated levels of CO2. The aim of the FACE system is to provide a level of CO2 enrichment 200 ppm above tmbient throughout the plot, without changing any other aspect of the microclimate within the plot. However, there can be substantial spatial variation in the CO2 level from the center to the edges of the plots. One of our main objectives was to predict the seasonal mean levels of CO2 for multiple subregions within the plots. However, the dataset was very large and followed a nonnormal distribution. Furthermore, the mean and variance were nonstationary. To overcome these difficulties daily means were used rather than individual measurements, the mean was modeled with multiple covariates that varied over both time and space, and the variance was modeled as an increasing function of the square of the distance from the center of the plot. A separable space-time covariance structure was used, and estimation was performed using nonlinear methods, REML, and EGLS. Finally, cross-validation was used to assess the validity of the model.

Key Words

EGLS Elevated CO2 FACE Nonlinear mixed models Nonstationary mean and variance REML Separable space-time covariance 


  1. Cressie, N. (1993), Statistics for Spatial Data, New York: Wiley.Google Scholar
  2. Cressie, N., and Huang, H. C. (1999), “Classes of Nonseparable, Spatio-Temporal Stationary Covariance Functions,” Journal of the American Statistical Association, 94, 1330–1340.MATHCrossRefMathSciNetGoogle Scholar
  3. Christensen, R. (2001), Advanced Linear Modeling: Multivariate Time Series, and Spatial Data: Nonparametric Regression and Response Surface Maximization (2nd ed.), New York: Springer.MATHGoogle Scholar
  4. Davidian, M., and Giltinan, D. M. (1995), Nonlinear Models for Repeated Measurement Data (1st ed.), London: Chapman & Hall.Google Scholar
  5. De Cesare, L., Myers, D. F., and Posa, D. (2001), “Product-Sum Covariance for Space-Time Modeling: An Environmental Application,” Environmetriec, 12, 11–23.CrossRefGoogle Scholar
  6. Gneiting, T (2002), “Nonseparable, Stationary Covariance Functions for Space-Time Data,” Journal of the American Statistical Association, 97, 590–600.MATHCrossRefMathSciNetGoogle Scholar
  7. Kuehl, R. O. (1994), Statistical Principles of Research Design and Analysis, Belmont, CA: Wadsworth.MATHGoogle Scholar
  8. Kyriakidis, P. C., and Journel, A. G. (1999), “Geostatistical Space-Time Models: A Review,” Mathematical Geology, 31, 651–684.MATHCrossRefMathSciNetGoogle Scholar
  9. Littel, R. C., Milliken, G. A., Stroup, W. W., and Wolfinger, R. D. (1996), SAS System for Mixed Models, Cary, NC: SAS Institute Inc.Google Scholar
  10. Mitchell, M. W. (2002), “Testing Separability of Spatio-Temporal Covariances,” Ph.D. thesis, North Carolina State University.Google Scholar
  11. Okada, M., Lieffering, M., Nakamura, H., Yoshimoto, M., Kim, H. Y., and Kobayashi, K. (2001), “Free-Air CO2 Enrichment (FACE) Using Pure CO2 Injection: System Design,” New Phytologist, 150, 251–260.CrossRefGoogle Scholar
  12. Posa, D. (1993), “A Simple Description of Spatial-Temporal Processes,” Computational Statistics and Data Analysis, 15, 425–437.MATHCrossRefMathSciNetGoogle Scholar
  13. Schott, J. (1997), Matrix Analysis for Statistics, New York: Wiley.MATHGoogle Scholar
  14. Stein, A., Van Groenigen, J. W., Jeger, M. J., Hoosbeek, M. R. (1998), “Space-Time Statistics for Environmental and Agricultural Related Phenomena,” Environmental and Ecological Statistics, 5, 155–172.CrossRefGoogle Scholar

Copyright information

© International Biometric Society 2003

Authors and Affiliations

  1. 1.BDResearch Triangle Park
  2. 2.North Carolina State UniversityRaleigh

Personalised recommendations