2010-08-27: Virtual Workshop WCS Implementation

From Earth Science Information Partners (ESIP)
Revision as of 08:27, September 8, 2010 by Erinmr (talk | contribs) (→‎Participants)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Main ESIP page >> Main AQ Work Group page >> Interoperability of Air Quality Data Systems | Standards Implementation Virtual Workshops

Telecon/Screenshare Info

1:30pm EDT
Phone: 866-489-0573 (US, Canada, VOIP); 205-354-0149 (international)
Meeting Code: *2279431*

Screenshare details:
Please join my meeting, https://www1.gotomeeting.com/join/954083041


Introduction

Who is this workshop for? Anyone interested in:

  • Making their data more widely accessible
  • Using standard tools and catalogs instead of custom-made
  • Need help adding a standard interface to data

Participants in this workshop may have varied backgrounds on web services (from novice to expert). The varied experience will allow participants to form mentor relationships.

Problem: Data is distributed, in different formats and can't be accessed by common tools within the AQ Community

Solution: Create a standard interface to the data using OGC Web Coverage Service

In-depth Details (programmers?)

Case studies of how Data Systems Implemented WCS:

How-To

Questions about WCS

What WCS questions/issues have you encountered? If you know the answer to these questions please respond below the question with ** and add your name.


Questions from Steve Ludewig, Sonoma Tech

  • The WCS protocol – how can/should it be extended for air quality
    • Units are the most glaring omission.

<units>ug/m3</units> RSIG includes units in the label: <label>pm25_daily_average(ug/m3)</label> -- Todd Plessel

    • Meeting notes: Connect to the WCS revision working group
    • Meeting notes: New major changes coming with WCS 2.0 - Comment period ended April 14, 2010.
      • Structure revamped
      • Concise document (core and extensions - 1 extension for making CF-netCDF)
      • Units are mentioned
      • GALEON paused until 2.0 came out, anticipating major changes
  • How should we handle different or even multiple parameter requests? (separate arguments? Lists as arguments?)
  • Duration codes?
    • WCS spec is a UTC time range, e.g.,: TIME=2006-07-05T00:00:00Z/2006-07-09T23:59:59Z so all WCS implementation should support this.
    • Meeting notes: PT1h, PT1d ... ISO Std.
  • Parameter occurrence codes (POC)?
    • More trouble than they're worth for most users who would rather just think about named variables over time and spatial ranges. (But feel free to extend your own implementation if needed to support your users.) -- Todd Plessel
  • Time_Sequence: Ours currently is a single date/time -- should it accommodate a time range (performance is a concern)? What time standard? (UTC?)
    • The GSFC WCS servers support time range and will pack multiple time slices (e.g., daily, monthly) data into one single file, thus increasing the dimensionality of the output, e.g., 2D spatial => 3D spatial and temporal. There might be, however, a performance issue. If a client requests a long time period, e.g., an entire year, for a daily product, the server may not be able to fulfill the request. Chris Lynnes suggested that the server only do one time step, which means that the client shall send multiple requests for multiple times and assemble the after receiving the data. --Wenli Yang
    • Yes. It must to be WCS-compliant. Absolutely otherwise it cannot be readily combined with other global data. WCS spec is a UTC time range, e.g.,: TIME=2006-07-05T00:00:00Z/2006-07-09T23:59:59Z

so all WCS implementation should support this. -- Todd Plessel

  • Use of an API key for authentication.
    • Authentication should not be required for a pubic WCS. (But customize yours if you must.) -- Todd Plessel
  • Meeting Notes: Discussion - Two perspectives 1. To implement standard compliant and then wrap the services to fit a particular project needs with a smart client 2. Serve the project first and if possible then add standard interface.
  • Meeting Notes: Discussion - No way to specify the limit that a client can request in the describe coverage (i.e. only data that can be returned in 1 min), so some providers aggregate and subset ahead of the web service.
  • Output (payload) formats
    • This is a crucial aspect for compatibility and performance that is not addressed by the WCS spec (presumably since it is not a data model). E.g., FORMAT=netcdf does not specify the conventions.RSIG supports the following output formats: http://badger.epa.gov/rsig/webscripts.html#output
    • GSFC servers currently support HDF-EOS and netCDF-CF1, but there are issues for some coverages in CF1 compatibility. --Wenli Yang
  • Field naming
    • CF Conventions may eventually become widespread, but they need to be extended for AQ variables (e.g., see CMAQ variable names). -- Todd Plessel
  • Meeting Notes: Discussion Need a data model
  • Meeting Notes: Discussion netCDF-CF conventions are proposed to OGC as an independent, binary encoding with core and extensions
  • CF Conventions need to be extended with AQ names CF Standard Names already submitted for Atmospheric Chem/Aerosol community

Implementation of WCS

Testing Services

Outcome

List your server here if you have a WCS implementation. Test with the Service Checker for GetCapabilites, DescribeCoverage and GetCoverage with the WCS Service Tester

Participants

Next to your name add if you are looking for guidance or have implemented a successful WCS and could provide guidance

  • Ben Domenico
  • Wenli Yang
  • Steve Ludewig
  • Rudy Husar
  • Steve Browdy
  • Stefan Falke
  • Erin Robinson
  • Carol Meyer
  • Karl Benedict
  • ...

Next Workshop

We will continue the discussion on WCS in telecon/webex form in 6 weeks. In hopes that the information gained from this session can be implemented/tested and we can move forward with a set of interoperable web services.