A Computable Spatio-temporal Framework For Solar-terrestrial Space

Scientific data is the core of solar-terrestrial space physics,whose scientific discoveries and business applications are dependent on the sharing,calculation and processing over these date.In addition to the big data characters,including high volume,multi sources and heterogeneity,the space physical data exhibits distinguished spatial and temporal features,such as sphere-shell spatial distribution,multi spatio-temporal coordinate systems and resolutions.Moreover,large amount of space physical researches aim to analyze the data’s spatio-temporal features.Applications like fine structure identification,coupling relationship research,spatio-temporal correlation and evolution analysis have been increasingly prominent.However,the prevalent data organization methods in this discipline are semantic-based,through which the data’s spaito-temporal features are ignored.Due to the mismatches between sematic-based organization and spatio-temporal oriented applications,data organization and application are usually performed by independent systems,which impedes the seamless connection between the two parts.This paper proposes a computable spatio-temporal framework for solar-terrestrial space,aiming to support integrated spatio-temporal data organization,storage and computation.The framework’s basic structure and hot issues that need to be solved urgently are overviewed in the paper also.The spatio-temporal data model and calculation operators in the framework are researched respectively: the data model should support for the spaito-temporal discretized data organization and distributed data storage;the operators are used to provide the spatio-temporal calculations of scientific data such as spatio-temporal data query,metric calculation and topological calculation.Based on the researches mentioned above,a prototype system of the framework is built to verify its validity and efficiency.The paper works around the following aspects:1.Conceptual analysis and design of the computable spatio-temporal framework for solar-terrestrial space.Based on the analysis of the spatial and temporal characteristics of the solar-terrestrial space physical data and the existing spatial and spatio-temporal data models,this paper put forward the design objectives,basic concepts and compositions of the framework,and divides the whole framework into two parts,Spatio-Temporal Data Model and Spatio-Temporal Calculation Operators.The design principle,research foundation and route of the data model are clarified,after which types of the operators supported by the framework are briefly summarized.2.Design HTM-ST,a spatio-temporal data model.Based on the spatial and temporal characteristics of the sphere-shell space distribution and multi spatio-temporal resolution,the spatio-temporal subdivision schema of HTM-ST data model is constructed by extending HTM’s spherical units to equal-divided time dimension.In addition,this paper designs and compares two coding methods,coupled coding and uncoupled coding,for HTM-ST’s grid cells.Meanwhile,combining with the coupled coding method and the data’s semantic characteristics,the spatio-temporal storage schema is designed in accordance with the RowKey storage rules of HBase.3.Design spatio-temporal calculation operators based on HTM-ST.Based on the HTM-ST spatio-temporal data model,a variety of general-purpose spatio-temporal calculation operators are designed to realize the basic computing support for the services in the upper level.In this paper,three kinds of commonly used operators are studied,including spatio-temporal data query,spatial metric calculation and spatial and temporal topological calculation.4.Prototype system construction and key technology verification.The prototype system of the computable spatio-temporal framework is designed and implemented based on HBase and existing solar-terrestrial space physical data resources.Several data service applications are constructed upon the HTM-ST data model and spatiotemporal calculation operators to verify the framework from the aspects of the data model,data query and global data interpolation based on the local time.