Premium
Power law decay in model predictability skill
Author(s) -
Chu Peter C.,
Ivanov Leonid M.,
Kantha Lakshmi H.,
Melnichenko Oleg V.,
Poberezhny Yuri A.
Publication year - 2002
Publication title -
geophysical research letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.007
H-Index - 273
eISSN - 1944-8007
pISSN - 0094-8276
DOI - 10.1029/2002gl014891
Subject(s) - predictability , term (time) , outlier , buoy , climatology , power law , probability density function , econometrics , statistical physics , function (biology) , meteorology , statistics , environmental science , mathematics , geology , physics , quantum mechanics , oceanography , evolutionary biology , biology
Ocean predictability skill is investigated using a Gulf of Mexico nowcast/forecast model. Power law scaling is found in the mean square error of displacement between drifting buoy and model trajectories (both at 50 m depth). The probability density function of the model valid prediction period (VPP) is asymmetric with a long and broad tail on the higher value side, which suggests long‐term predictability. The calculations demonstrate that the long‐term (extreme long such as 50–60 day) predictability is not an "outlier" and shares the same statistical properties as the short‐term predictions.