Past climate changes have caused certain perturbations in the subsurface temperature field. Their inversion yields the time variations of the ground surface temperature (GST), the solution of which however may be highly non-unique. To incorporate additional information on the behaviour of the unknown function usually helps treat the non-uniqueness. The traditionally used information imposing bounding and smoothing constraints on climate history can be further significantly extended by including stochastic properties of climate. It was shown that climatic records reveal: (1) long-term persistence of climate, and (2) high intermittency of climate variations characterized by hyperbolic decrease of probability tails. We have quantified and incorporated such information into the inversion method in the form of the covariance matrix of the unknown parameters. Inversion of almost 100 temperature-depth profiles from the Czech Republic (Bodri and Cermak, 1995, 1997) provided more meaningful climatic histories in comparison with simple inversion and increased the time span that could be resolved. In present work the effectiveness of the method is illistrated on the examples of 5 selected temperature logs.
ASJC Scopus subject areas
- Earth and Planetary Sciences(all)