NASA GIBS/Worldview Visualization – Granules / Vectors / Curtains


NASA’s evolving Global Imagery Browse Services (GIBS) project provides raster imagery for a wide range of geophysical parameters across EOSDIS missions. This imagery is available through a set of web services and associated standard interfaces which facilitate efficient transmission of raster layers to mapping clients which consume geospatial imagery for a wide spectrum of applications.  The varied types of Earth science data within the ever-growing holdings of EOSDIS require GIBS to expand and evolve beyond its current operational focus of serving daily global and polar raster composites.

This ESIP session focuses on presenting current and planned GIBS development activities for the purpose of soliciting comments, questions, and user scenarios. Topics include visualization support for:

  • Sub-Daily Data (per science product file/granule) - Visualization of individual data files, which typically have a temporal coverage that is less than one day.  Common use cases include visualizing single swaths/granules/scenes.
  • Vector-Based Data – Visualization of vector-based data in non-raster formats.  Common use cases include client-side styling, efficient tiling scheme’s, etc.  Considered vector products include orbit/ship tracks, dense data products, and gridded (varying by zoom level) products.
  • Curtains/Vertical Profiles – Visualization of non lat/lon data products that are best visualized along a “z axis”, often associated with elevation or pressure level.  
  • Interactive Analysis (R&D work) - Data visualization techniques facilitating direct interaction with the source science data behind the GIBS imagery, locally within a browser or science tool.

NASA GIBS/Worldview Visualization
Granules, Vectors, Curtains, & Dynamic Data
TIROS-1 Launched in 1960
Nimbus-1 in 1964
Now Today – 20 active missions
            Better imagery, and better data
            Quantifiable science products -> long-term analyses
            Variable data types, levels, structures
            Massive Data Volume (15 PB of data, 16TB/day)
Global Imagery Browse Services (GIBS)
            Transform how users interact and discover NASA Earth data
Ingest processes supported by TIE
Tile pyramid is through Meta Raster Format (MRF)
OnEarth provides access to that
Worldview allows for visualization
Sub-Daily Imagery
            Visualization of individual data files
            Temporal period is less than a day
Storage through Enhanced Meta Raster Format (MRF)
            Web Mapping Tile Service (WMTS)
            Web Mapping Service (WMS) supported
            File-base delivery
Example P.O.C. Demonstration
Vector Base Data
Vector – a geographic feature that can be represented by a point, line, or polygon.
            – a quantity with magnitude and direction
Samples – Orbit Tracks, Earthquakes, Windspeed, Etc
Currently, everything is served as rasterized image tiles.
            This is not ideal for several reasons (aesthetics, more work, etc).
Proposed solution:
            Store vector-based datasets using Esri Shapefiles, and MRF (Meta Raster Format)
            Clients request data as: Mapbox Vector Tiles, GeoJSON, and PNG
Mapbox Vector Tiles
            Optimized for better performance, though there is slight precision loss.
            Binary (not human readable).
            Takes longer to load, but better precision
Rasterized Tiles
            Pre-generated and served via WMTS
There are many ways of access
Planned support:
            WMTS and WMS for vector tiles and rasterized vectors
            WFS for GeoJSON
Various server configurations have been tested. These all are different, but many struggle with treating time varying datasets.
Styling – Clients can do their own styling. Guidelines from this should come from data providers.
Workflow is demonstrated in the slides.
For data providers:
            There are a variety of needs including metadata and others.
Data consumers:
            Vectors and PNGs available based on user needs.
            Some datasets are better visualized as vector, even if they are raster.
                        Good point, sparse datasets that could be represented as points can be stored as vectors for space, styled differently, and clustered.
            What is the plan for the NASA aircraft campaigns?
                        Our vector data could handle this. The subdaily technology discussed before is also potentially capable of dealing with these sort of things. However, there are challenges still to be tackled, including metadata. Necessitates 3D visualization.
Curtain/Profile Data Visualizations           
An ultimate goal is to produce vertical profiling.
These data exist and are important.
            However, X/Y Tiling does not apply. Additionally, this sort of work has been completed before.
            How do you serve these images?
            How is the ground track AND image provided together?
            How do you provide vertical axis values?
            Unknown unknowns.
Introductory discussion here at Summer ESIP.
GIBS-led Technical Working Group

Goal:    Develop a list of use cases to guide developmenti.Develop a list of existing data
ii.Identify stakeholders
More discussion here:
Dynamic Data Visualizations
Allow for more interactivity and analysis of data within browser.
Currently serving only Pre-Generated Imagery
            Limited ranges on data
            Considerable storage requirement issues
            Quickly view raw data values (hover over, etc)
            Change color scales and palettes
            Perform quick statistical analyses
Level 2 Swath Dynamic Visualization
            Data are arranged in “instrument space”
            Stored in granule files
            No resampling or interpolation within the grid
Example: AIRS surface temperature
How then do you create a “science image”?
            Calculate the actual shape of each footprint using known geometry
L2 Dynamic Data Visualization in GIBS
Current Status
            Several Activities including:
                        Created index-to0MRF process for AIRS L2 Data
                        Creating IDL client
Additional questions
            These vary based across Data Access, Index Image Generation, and Client-side visualization.
Level 3/4 Swath Dynamic Visualization
Image encoding approaches
            24-bit data PNGS (pack data values into 3 8-bit channels of a PNG image)
Limited Error Raster Compressions (LERC)
            ESRI format
            Not handled natively in browser
Flow map is shown.
Current Status
            Encoded PNGs and LERC times are working in OnEarth
            Test clients are available for PNGs
            Evaluating LERC
            Goal: Basic proof-of-concept by October

Cechini, M.; Roberts, J.; Thompson, C.; Boller, R.; NASA GIBS/Worldview Visualization – Granules / Vectors / Curtains; 2016 ESIP Summer Meeting. ESIP Commons , April 2016