The summary is not yet available. Stay tuned.
Kevin Tyle, University of Albany, Chair
Michael Baldwin, Purdue University
Martin Baxter, Central Michigan University
Sen Chiao, San Jose State University
Jennifer Collins, University of South Florida
Ibrahim Demir, University of Iowa
Bart Geerts, University of Wyoming
Kevin Goebbert, Valparaiso University
Steve Lazarus, Florida Institute of Technology (remote)
Gretchen Mullendore, University of North Dakota
Sam Ng, Metropolitan State University of Denver
Russ Schumacher, Colorado State University
Kimberley Hoogewind, Purdue University
Becky Cosgrove (CONDUIT)
David Plummer for Michelle Mainelli (GEMPAK-NAWIPS/AWIPS II) (remote)
UPC Staff Attending
| John Caron
| Ryan May
Introduction: New Members
New members were welcomed to the committee and provided a brief introduction. Dr. Ibrahim Demir (University of Iowa) is an assistant research professor specializing in informatics and works within the hydrology community, Dr. Kevin Goebbert (Valparaiso University) is an assistant professor of atmospheric science with a strong interest in using Python for visualization and Dr. Gretchen Mullendore (University of North Dakota) is an associate professor of atmospheric science that studies effects of pollutants on the upper atmosphere during severe storms. This meeting will be the final session for Dr. Marty Baxter and Dr. Jennifer Collins, and Dr. Bart Geerts as User Committee members. Their participation and contributions were applauded.
Date for Spring meeting:
The next meeting will take place on Thursday, March 26th and Friday, March 27th. Kevin Tyle expressed an interest in coordinating this meeting with William Gallus the chair of the Strategic Advisory Committee. However, since the spring SAC meeting often takes place somewhere other than Boulder, a joint meeting may not be possible until the fall.
Review of Action Items
Spring meeting summary is available here. It was noted that this was the first meeting for the new NASA representative, Christopher Lynnes and then he highlighted www.earthdata.nasa.gov which may be of interest for User Committee members.
A brief discussion subsequent conversation noted that the AWIPS II development guide is still pending. Scott Jacobs had committed to provide the guide by the following week.
Questions and discussion following the Director’s Report:
There was a conversation around issues associated with large data volumes from things like HRRR. Mohan noted that there are not capacity issues on the UPC side but university users may have a hard time handling the large volume of data. THREDDS was suggested as a way to handle this traffic. Comments on this suggestion, noted that most people pull a small amount of data and the associated issue is having servers that can process the requests. Rich Signell noted that his organization previously used LDM but now they go through Motherlode.
Mohan noted that NOAAPORT will be doubling its throughput capacity by the end of the month (by October 1st); however, this requires adjustments to the satellite dish used by the UPC which is owned by NOAA NWS and located at UCAR for COMET education activities. A request has been submitted to update this dish to address the pending NOAAPORT changes.
In a brief conversation on cloud-related activities, Kevin Tyle noted that he had been able to install both EDEX and visualization on a single workstation; however, together, they had been too much for his hardware. When he used the Azure cloud-hosted EDEX server, visualization on the local machine performed adequately.
Programmatically, there was a discussion of the ideal size for Unidata and whether non-core awards were worthwhile. Mohan noted that we have to pool resources and we think non-core awards are valuable. There are 8 or 9 non-core awards, which make things like Rosetta or DAP4 available. However, there are challenges with fractional tasks for people that canâ€™t be ignored.
Sam Ng: Noted that he has ordered a new server and graphics card. He was able to get AWIPS II up and running but does not currently have any data. He noted that having a 2G graphics card has really improved the IDV speed. Additional comments included that some students are still using GEMPAK and that they were able to supplement their equipment award with university funds.
Marty Baxter: We finally got our 16 screen map wall up and running, and the content is also mirrored on the web. I have a CAVE client running on my office machine and hope to have a program machine running AWIPS II. Our weather station is up and reporting to weather underground and I would like to work with Sean Arms to get it in ROSETTA. I would also be very interested in working with the UPC to get a touchscreen version of IDV working. I also installed THREDDS and OPENDaP is really helpful but there are some quirks.
Kevin Goebbert: I continue to use Unidata products. My freshmen still start with GEMPAk but I'm introducing python to the Juniors and Seniors.
Ibrahmi Demir: I use the LDM to access data, and now we are working on a cloud installation of LDM. I'm also working on a web-based visualization system. We also have a UPC grant for AWIPS II.
Jennifer Collins: I'm still on sabbatical; however, I have one student using IDV for a project and I'm working on the announcement for the regional workshop in April 2015 that will be combined with the West Central Florida AMS meeting.
Russ Schmacher: We received an equipment grant a year ago, students were finding the 50 different copies of data frustrating and were asking for a central place for data sets. The new server is much faster than the previous work flow. We have a RAMADDA set up but I could use some help using it. Brian Maps has been doing a lot with the IDV, and we just had a proposal on the IDV and WRF recommended for funding from NASA.
Kevin Tyle: Busy with David Knight working on transition to University data center, we have been playing with virtual machines and installing LDM on those platforms. RAMADDA has been running for a while with increased buy in from faculty and students. I'm trying to get the word out to leverage OPENDaP, which is starting to be used by both faculty and graduate students.
Rich Signell: Our local USGS group is running an ocean forecast model, accessed through Motherlode instead of LDM. We convert outputs into netcdf and put it on TDS for users to access. We have had a hard time translating discrete sampling data and serving out observational data. A best practices document for observational data would be great.
Sen Chiao: My students use IDV a lot for the weather briefing course and I'm in the process of setting up AWIPS II for the program. This involves a new machine and renovating a room. We started digging into netcdf python plus OPENDAP to plot wavewatch 3. We are ingesting HRRR data and also recently got NAM12 into our system.
Michael Baldwin: Everything has been centralized, though we do still have an IT person in the department (who is attending the LDM training next month). At this point no one really understands what is happening with the data feed at Purdue. Trying to incorporate more and more Python. Lots of personnel leaving.
Kimberley Hoogewind: I'm not teaching this semester since I'm focused on research; running lots of models and generating lots of data. I'm trying to get concurrent processing to increase the number of runs.
Bart Geerts: I am mostly in maintenance mode; did migrate to central IT on servers, same person managing LDM, recently developed a tool with Greg Stossmeister to generate a tool to for web-based display of aircraft data on mobile phones. We have added HRRR data to surface products.
Gretchen Mullendore: I work extensively with netcdf, and are starting to work with Python. Some data flow and distribution challenges within our school. We use IDV and gempak but they only work well on certain machines. There is resistance to OPENDAP because people feel like it is too slow unless they download locally. Sam Ng noted that he had similar issues which were resolved with a video card improvement.
Steven Lazarus: I'm still running a display monitor, on a major thoroughfare on campus, we still run the LDM, working with Jennifer Collins on the spring workshop, spending a lot of time working with the IDV in the classroom and I owe Julien Chastang an IDV video.
Michael James noted that due to technical issues with his new AWIPS II laptop he would not be running a demonstration. There was a question whether RTOFS (Real Time Ocean Forecast System) was showing up yet, to which Michael noted that one version is available but ESTOFS (Extratropical Surge and Tide Operational Forecasting System) has not been integrated yet. UFRAME was mentioned as the open source underlying solution used by Raytheon.
The group asked Michael how many active beta testers were working with AWIPS II. He noted that there are 50 users on the mailing list from 18 different educational institutions. Michael noted that he cannot track who has downloaded from your site but he can identify who has accessed our server (8 universities). There is also a lot of .com activity on the user list. We get a lot of helpful feedback from .com users but he have to be careful with support for .com.
Committee members asked if there a User Guide? At this time there is not one currently available and it is behind schedule for the beta testing. A lot of the user guide type of information that Michael has is not cleared for release or is no longer valid with the Unidata release. Committee members noted that they can get AWIPS II working but they donâ€™t know what to do with it once it is up.
Kevin Tyle ran into a couple of errors trying to uninstall some RPMs.
Michael James noted that he would like to issue a new release but also has to keep it stable through next monthâ€™s training.
Upon being asked, half of the User Committee members identified as actively testing AWIPS II.
All noted that documentation for AWIPS II would be helpful.
When asked, committee members specified that they do not want to modify the Java code.
Lastly, around 30 WFOs have transitioned to AWIPS II, next two groups have 16 each. Everyone at the local offices will be on AWIPS II before the current contract ends.
David's slides. (Presented remotely via WebEx.)
The Boulder facilities will be updated Fall 2015; however, we are still ready to add data? Yes.
During the potential 2015 CONDUIT Survey, there should be a question about push vs pull. For HRRR, do we prefer operational or research with less reliability. GSD feed is down about 10% of the time.
Could Jumbo THREDDS support both versions of HRRR? Yes, until we bring in HiWIPP.
Is there Interest in .25 deg GFS? Yes (.25 deg global, 20km regional).
Committee members also request for 4km NAM nest to be available for CONUS.
Committee members asked about the location for up to date model plans? Becky noted that this is not currently publicized but that there is a quarterly update to the Director, Becky will check if it can be shared.
Committee members need a mechanism to report issues with NOAAPORT. For now, Becky noted that they should send such issues to her.
Staff status reports were made available to the committee prior to the meeting. Committee members had comments or questions on a subset of the projects reported upon:
Question raised whether cloud is anything off-site or is more focused on data access and analysis? No clear consensus answer to that question. Rich Signell noted that they are experimenting with enterprise Wakari accounts at USGS to allow folks closer access to the data/server side processing. Additional question if any universities setting up clouds? UMASS Consortium with cloud resource, CSU has a cloud, and University of Iowa has one at the data center.
There was a question about PyCDM, to which Rich Signell stated that IRIS has the CDM stuff in it, and it would be nice as a component instead of all of the IRIS stuff. Rich Signell noted that Rich Haddersley says they do not have a cross variable CDM in IRIS. IRIS didn't work on PCs because UDunits didn't work on windows. There was a nice synergy where making UDunits work on windows resulted in a whole new set of capabilities. There was a separate conversation about the maturity of Cartopy. The consensus is that it is still evolving but definitely going to replace basemap. Cartopy is immature regarding documentation; there are a few quirks to resolve.
Question about a metric on response time for support. Answer is no, it varies by program. Follow-up question whether anyone monitors stackoverflow? The answer is yes, but timing varies. For example, with 6 people tracking THREDDS, it may take 6 weeks for the expert to get to the question. The bulk of the stackoverflow questions are netcdf. Folks are welcome to submit through the Unidata website, stack overflow, or github. Each approach has their own strength. Github is better for developers, esupport is better for conversations, jira for long term problems.
Steering Team meeting a couple of months ago, could use more members, some of it is just learning what is new. Concern about development of WMS, concern about the robustness of TDS outside of UPC, started plotting observations, IDV tries to read BUFR, what is the level of support for BUFR format. User Committee members commented that the new list that goes to everyone is good. The list all users is very helpful. In general, Christian is working on point development, so hopefully less backlog
Unidata Outreach Progress
Comprehensive Metrics Data
The presentation on the Triennial Workshop is available here.
Comments and questions included:
There is some value in featuring at least one break out where students work as a team to give feedback.
Perhaps at least one collaborative session to work on new skills with your own data sets at the end of the week.
There was a suggestion to have an enforced faculty/student combo; however, the discussion concluded that while it might be powerful it would be difficult in practice.
One potential carry-over project that could continue after the event is a product/guide focused on data management best practices and integration with a full spectrum of tools that would serve as a resource for the community.
Another possible exercise for participants would be a Rosetta session converting your dataset (in any format) and leave with netcdf.
Rich Signell's presentation, as the 2014 DeSouza award recipient, is available on the Unidata Seminar Series page.---
This presentation is available here.
Committee members expressed no concerns regarding these proposed questions so they will serve as the template. Members should feel free to deviate from this template based on the flow of conversation.
Jeff Weber provided an update and demonstration of the Teaching Resources product. Committee members noted some confusion that the COMET case studies were included primarily to demonstrate the functionality of the Teaching Resources tool and this was not focused on those case studies. They noted that the demonstration was helpful and that the product could be very meaningful with further documentation and user guides such as screencasts demonstrating steps for uploading.
There was some conversation about the ability to classify existing products (i.e. case studies) as a Teaching Resource. Currently, this requires a new entry. However, Jeff will work to make this possible and largely seamless.
Ward's presentation is available here
The chair chose to first ask committee members for open-ended blue sky requests. Responses are below:
Kimbereley Hoogewind: Noted that newer versions of Gempak do not support GARP. GARP offered cross-sections which is missed but members noted that GARP did not support modern items.
Michael Baldwin: Noted a need for more time, which could be helped by decreasing the amount of time necessary to learn all of the relevant tools.
Gretchen Mullendore: Added that data collected from automobiles, possibly submitted to RAMADDA, could be of interest.
Sen Chiao: Is there a way to create tutorials on these products or services, or increase the ease of use. Minimize the learning curve as much as possible. I would like to create a website that interacts with IDV. (Yuan Ho noted that the requested function can be done with RAMADDA).
Rus Schumacher: Development of software tools to look forward to ensemble forecasting, IDV is doing some of that by allowing the creation of your ensemble. Within AWIPS II, that seems to be a lower priority. Looking for model comparisons all of this info is available but it requires going to different sites.
Rich Signell: Would like the functionality of the IDV in a more easily modifiable language such as Python. That would require CDM feature types in python. Maybe there could be something like the function working with Matlab within larger python construct. The emphasis on building in the community development would be able to reduce the burden. Unidata could take a lead for best practices on creating dynamic web content for atmospheric sciences.
Sam Ng: Suggested ny tutorials related to RAMADDA would be very helpful.
Marty Baxter: More activity related to training and less focused on specific services, perhaps add an FTE for training and outreach. More of an advocacy type role, suggesting that this needs to be incorporated in the curriculum. UCAR membership committee could be an avenue for communicating the need for curriculum updates. The AMS curricula group could be another avenue.
Steven Lazarus: IDV data sources can be difficult to navigate. It would be nice to have access to more data through the catalogues (such as polar orbital data). A lot of satellite data sources are redundant.
Kevin Goebbert: GEMPAK has a limited shelf life, having that replaced with something that easy to use with quality graphics is essential. Batch processing would be great.
Ibrahim Demir: More that would move away from desktop because that is going away. Many colleagues are using LDM but they are installed by Sys Adm that cannot optimize configuration maybe one way to address that limitation is to move to the web. New libraries like WebCL allow you to write native code for visualizing large scale data on the client side.
Gretchen Mullendore: Some benchmark or guide on performance to know that we are suboptimal and can ask for help.
Jennifer Collins: Concur with MaBa sentiments; special issue on teaching resources in a journal. Diversity on the committee to include more full professors.
Kevin Tyle: I'm interested in trying to identify some more dedicated computer support folks to participate in this group. I'm also interested in products that leverage WMS.
Ryan's presentation is available here.
Actions will be compiled and the meeting summary will be drafted by Kevin Tyle and Josh Young for circulation shortly.
Steve Lazarus and Michael Baldwin agreed to serve as co-chairs for the 2015 Users Workshop.