Nonparametric Smoothing of Irregularly-Spaced Palaeo Records: Trend Estimation with Application to Climate Change

• Activity Number: 512
• Type: Contributed
• Date/Time: Thursday, August 2, 2007 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics and the Environment
• Author(s): Patricia Menendez*+ and Sucharita Ghosh
• Companies: Swiss Federal Institute for Forest Snow and Landscape Research WSL and Swiss Federal Institute for Forest Snow and Landscape Research WSL
• Keywords: Palaeo time Series ; Climate Change ; Irregularly spaced data ; Kernel smoothing ; Long-range dependence ; Trend estimation

Abstract

Palaeo data such as fossil pollen or ice isotope records are irregularly spaced in time due to the nature of the dating procedure. Such studies consider the evolution of the processes along the age scale. In statistics, uneven and random sampling schemes in time series data have been considered by various authors, in particular for alias-free spectral estimation (e.g., Shapiro and Silverman 1960, Masry 1978 and others). We consider the nonparametric smoothing problem, when the data have been sampled irregularly and randomly in time. As for the model, of special interest are processes with time varying distribution functions, such as a time dependent one-dimensional transformation of a stationary Gaussian process; see Ghosh, Beran and Innes 1997 and Ghosh and Draghicescu 2002. Asymptotic results under long-range and short-range dependence are derived with examples of fossil records.