This is the main page for the OSSE project. OOI CI will conduct an Observing System Simulation Experiment (OSSE) in November 2009 to prototype the designed capabilities of the OOI CI to support field efforts in a distributed ocean observatory in the Mid-Atlantic Bight. The goal is to provide a real oceanographic test bed in which the designed CI technologies will support field operations of ships and mobile platforms, aggregate data from fixed platforms, shore-based radars, and satellites and offer these data streams to data assimilative forecast models. The MAB region is selected because of the existing communities and the presence of NOAA, ONR. The effort is coordinated by Oscar Schofield of Rutgers in the context of the MARCOOS effort. Planning and design of the experiment will occur in early/mid 2009. A virtual deployment occurred in September 2009. The field deployment experiment itself is scheduled for November 2-13 2009.
The core OSSE integration consists of four main activities:
- ROMS numerical model integration (group of Yi Chao, JPL)
- CASPER/ASPEN mission planning and control (group of Steve Chien, JPL)
- MOOSDB, MOOS-IvP autonomous vehicle control (group of Arjuna Balasuriya, MIT)
- Glider Simulator Environment and Field Deployment in Mid Atlantic Bight (group of Oscar Schofield, Rutgers)
An integrated Statement of Work (SoW) with a high-level design graphic can be found here. The OSSE group met April 28-29 at Rutgers University for the OSSE and Modelers Meeting. It was the kick-off face-to-face meeting for the OSSE group as well as a meeting with numerical modelers of the mid-Atlantic region.
The OOI CI Kick-Off meeting presented the OSSE as opening showcase for advanced science and technology. The one hour presentation in the plenary can be found on the meeting pageOSSE Plenary Slides for OOI Kickoff Meeting (Science Visions). There was also a detailed one hour demo. The meeting took place September 9-11, 2009 near Denver, CO.
RECENT: The OSSE Field Deployment occured Nov 2-13 2009 in the mid-Atlantic Bight, with the operations center at Rutgers University, a field station in Tuckerton NJ and ship time for the week of Nov 2. It will bring together six operational numerical models, observational data from satellite, radar and gliders, 4 AUVs of different type, 5 gliders (2 of which are taskable with advanced shore-side operations support). Results are collected here (OSSE Rocks!!) and will be incorporated into a report by the end of the year
- Field Deployment Coordination Page
- Field Deployment Results Collection Page
- OSSE Data Portal (at JPL, based on OurOceanPortal)
- OSSE Deployment Blog and interactive web page (at Rutgers)
- OSSE Numerical Model Integration coordination page
- OSSE Technology Integration coordination page
The figure below shows the main components within the OSSE and interfaces between the components.
Detailed component descriptions:
with scenarios, data flow and activities here.
- MOOSIvP Public (read-only) subversion repository at MIT: "svn co http://oceanai.mit.edu/svn/moos-ivp-public/trunk/ moos-ivp"
- Glider Data Example
Efforts continue to speed up as we prepare for the Fall/Winter OSSE demo for the OOI. We had a good call yesterday for the software planning and prosecution efforts, and wanted to highlight just a few things. For this effort, I remind everyone that the goal is to demonstrate integration of the subsystems (models, software, gliders etc.), so the don't stress about using the most newly evolving model (ie that will be improving forever, so just need model fields to test). If you have not gotten your information to Yi please do so this week. Also if your model is assimilates any data can send me the times over the day when the data is ingested. For some there has been discussion about resources, I have follow-up calls with Scripps this week, to work out some of those issues. Please send me a recap of (reasonable) needs. What is exciting is that OOI was approved by the OOI, and effective start dates for the cyberinfrastructure is this Fall! Some people suggested to interpolate the N model solutions onto a common grid over the field experiment domain. This way, we don't need to reveal which field is derived from which model, and a common ensemble mean can be calculated, and the associated error can be estimated. The JPL glider planning team is exploring ways to use this error information in the glider planning. Please let us know what do you think about this approach. Any other ideas and suggestions are greatly appreciated. We thank everyone for their efforts and are excited for the fall. New flowcharts and resources are posted on the confluence sit, feel free to use it as a community resource!