Data Rods: Managing Cryospheric Remote Sensing Data Entirely Within a Pure-Object Database

Climate data analysis frequently seeks to detect changes over time, yet much of the remote sensing data at the National Snow and Ice Data Center (NSIDC) is stored as separate geo-referenced files for single times. This organization becomes cumbersome as the data sets grow, requiring investigators to iteratively obtain and subset data files to the area of spatial and temporal interest. The Data Rods project proposes to prototype a new data organization, “data rods”, that will permit rapid, selective retrieval of cryospheric data across massive multi-modality data sets. The core concept is to reorganize these spatially arranged data sets into a multi-dimensional pure-object database wherein time is the Z component and sensor values add further dimensionality. A “data rod” is then a vertical data construct, a collection of all known values at a single location through time. A friendly web-based user interface will provide on-the-fly query construction and filtering across any selected set of data rods, permitting rapid, flexible retrieval of cryospheric data. The integration of useful algorithmic tools into the UI may enable new methods of visualization, analysis and data correlation across data sets. Our initial target area is the Greenland ice sheet. If successful, the project may be expanded to cover other cryospheric regions. Data Rods Conceptual Design