STARE INTO THE FUTURE OF GEODATA INTEGRATIVE ANALYSIS

Area of Science:

• Earth Science
• Data Integration
• Geospatial Analysis

Background:

• Diverse Earth observation data (e.g., visible-infrared imagery, radar) have unique strengths for scientific modeling.
• Integrating these varied data types, especially irregular low-level data, is currently a complex and non-scalable process.
• Existing data organization methods using rectilinear partitioning and integer indices hinder seamless integration.

Purpose of the Study:

• To address the challenges in scalably integrating diverse Earth Science data.
• To introduce a novel encoding method that overcomes limitations of traditional data partitioning.
• To enable more effective use of multi-source observational data in scientific research.

Main Methods:

• The study proposes the SpatioTemporal Adaptive Resolution Encoding (STARE) as an alternative to traditional geo-partitioning.
• STARE provides a unified encoding for geo-spatiotemporal location and neighborhood.
• This method transcends file-based structures and native array indexing.

Main Results:

• STARE enables diverse data to be organized on scalable, distributed computing and storage platforms.
• It overcomes the limitations of rectilinear partitioning and floating-point longitude-latitude (lon-lat) encoding.
• Facilitates integration of heterogeneous Earth Science datasets.

Conclusions:

• The SpatioTemporal Adaptive Resolution Encoding (STARE) offers a scalable and effective solution for Earth Science data integration.
• This approach enhances the ability to combine and analyze data from various sources.
• STARE paves the way for more robust scientific modeling and hypothesis testing.