摘要: Spatio-temporal semantics based on "object views" or "event views" has few abilities to represent and model the continuity and gradual oceanic phenomena or objects, which seriously limits the specific marine applications and knowledge discovery and data mining, so this paper proposes a hierarchical abstraction semantics with "marine spatio-temporal process→life span phases→evolution sequences→state units" and process objects included by level with "marine process objects→phase objects→sequence object→state objects" with the oceanic process characteristics into the marine process semantics. In addition, this paper designs the storage and representation of marine process objects using the backus normal forms (BNF) and abstract data type (ADT). Base on El Niño Southern Oscilation (ENSO) index and Chinese rain gauging station data, this paper also gives a case of study. The spatio-temporal analysis between ENSO process and Chinese rainfall anomalies shows that the marine spatio-temporal semantics not only can illustrate the spatial distribution of Chinese rainfall anomalies in different time scales at ENSO process, life span phases and state units, but also analyze the dynamic changes of Chinese rainfall anomalies in different life span phases or state units within ENSO evolution.