Interactive Comment on '' Long-Memory Processes in Global Ozone and Temperature Variations'' by C. Varotsos and D. Kirk-Davidoff. Rust, H. Atmospheric Chemistry and Physics Discussions, 6:S1182-S1185, 2006.
abstract   bibtex   
Long memory or long-range dependent processes are frequently claimed to underly the dynamics of observed time series. Many of the works reaching those claims are based on detrended fluctuation analysis (DFA) or related scaling analysis techniques. The general procedure of these heuristic approaches is to describe the behaviour of a certain characteristic with a power-law. Although this is surely a useful ansatz in many respects, its validity should be checked for every application. Especially, when two or more power-law relations in a different range of scales are used to describe the result. Despite the difficulties to interpret, a power-law is frequently preferred over more complex concepts allowing a physical interpretation because of the small number of parameters. When using multiple power-laws for different '' scaling regions'', this argument is corrupted.
@article{rustInteractiveCommentLongmemory2006,
  title = {Interactive Comment on '' {{Long}}-Memory Processes in Global Ozone and Temperature Variations'' by {{C}}. {{Varotsos}} and {{D}}. {{Kirk}}-{{Davidoff}}},
  author = {Rust, H.},
  year = {2006},
  volume = {6},
  pages = {S1182-S1185},
  issn = {1680-7375},
  abstract = {Long memory or long-range dependent processes are frequently claimed to underly the dynamics of observed time series. Many of the works reaching those claims are based on detrended fluctuation analysis (DFA) or related scaling analysis techniques.  The general procedure of these heuristic approaches is to describe the behaviour of a certain characteristic with a power-law. Although this is surely a useful ansatz in many respects, its validity should be checked for every application. Especially, when two or more power-law relations in a different range of scales are used to describe the result. Despite the difficulties to interpret, a power-law is frequently preferred over more complex concepts allowing a physical interpretation because of the small number of parameters. When using multiple power-laws for different '' scaling regions'', this argument is corrupted.},
  journal = {Atmospheric Chemistry and Physics Discussions},
  keywords = {*imported-from-citeulike-INRMM,~INRMM-MiD:c-13317395,autoregressive-model,modelling-uncertainty,power-law,regression,statistics,uncertainty},
  lccn = {INRMM-MiD:c-13317395}
}

Downloads: 0