The Estimation of Mesoscale Ocean Eddies Change Based on CBR

Case-based reasoning (CBR) method has been widely used to study geographical problems during the past two decades. However it is still not perfect particularly when employed to solve complicated geographical problems. Urgently needed improvement includes development of geographic data representation modeling and design of algorithm for spatial similarity computation and reasoning. This paper reports an improved CBR-based method for studying the spatially and temporally complex Mesoscale Ocean Eddies (MOEs). After summarizes the basic advantages and challenges of current existing quantitative methods, the paper first proposes that CBR approach, with support of GIS, can be employed to study variation of MOEs. Representation model was constructed to describe the case, i.e., MOEs. This paper then provides an algorithm to retrieve the inherent spatial relationships among cases, as well as a CBR similarity reasoning algorithm to predict change of MOEs. The method was finally tested by examining the MOE in the South China Sea and yields an average estimation accuracy of 80%. In summary, the CBR-based approach proposed in this study provides an effective and explicit solution to quantitatively analyze and predict the change of MOEs.