Witness the Arctic (WTA) had the opportunity to talk with Mark Serreze of the National Snow and Ice Data Center (NSIDC; http://nsidc.org/) and Jim Moore of the National Center for Atmospheric Research (NCAR; http://ncar.ucar.edu/) who lead the Advanced Cooperative Arctic Data and Information Service (ACADIS) team. We asked each of them a few questions about the new data management service. Here are excerpts from those conversations.
(Further information about ACADIS is available in this issue of Witness.)
WTA: What are some of the highlights of the new ACADIS system?
Serreze: This level of data management is relatively new for NSF. We have an opportunity now to start providing some services that can really benefit the community.
One of the services we will provide is a range of value-added products, in other words, taking the datasets out there and combining them in ways that will serve a much broader community. An example might be for the borehole data from arctic monitoring permafrost temperatures—our job would be to bring all those disparate datasets together so that investigators can make greater use of them. Which of these value-added products we need to focus on is something to be determined in collaboration with the ACADIS Advisory Committee and NSF management.
Moore: The fundamental initial improvement is to expand on the capabilities of the Cooperative Arctic Data and Information System (CADIS) (http://www.aoncadis.org/) to handle different data and formats from all the different types of programs funded by NSF's Arctic Sciences Division. This requires some improvement to the user interface that we use for creating metadata, for uploading datasets, and to improve visualization capabilities.
The user interface allows principal investigators to build the metadata file for each and every dataset they have and provides a way for them to upload the data into the archive so we can get it into a permanent facility. The website offers any user the chance to come in and look for the data they're interested in and then download the data directly to their computer.
Another opportunity here is with the development of a standardized metadata scheme that matches international standards. In matching the standards used by global data archive centers, we are going to open the opportunity to share data across archive sites. We have done some trials already with projects in Germany, Norway, and the U.K. The hope is that in the long run, arctic scientists can sit down at the ACADIS user interface and find data they are looking for from any of the major archive sites around the world.
WTA: What are you currently working on and what are the next steps for ACADIS?
Serreze: The next step is to transition to ACADIS Version 1, which is close on the horizon. In the meantime, any investigator who needs to place data with us can do so. One of the things we need to be determining fairly quickly is the level of service for each dataset. Each dataset doesn't necessarily get the same level of service—it depends on things like the perceived value of the data and how much we think it is going to be used. The high-value, integrated datasets would get a high level of service, while other datasets may get less. We have to come up with a good system to deal with that right off the bat. We're going to need very open communication. Those are the types of things we will be deciding with the principal investigators and also our Advisory Committee.
Moore: There is no higher priority than making sure the Arctic Observing Network (AON) community can continue to get their data and metadata into the archive. We've got to make sure we don't take a step backwards and that the tool continues to be operational—and for some projects that's easier said than done. It's one thing to develop a new tool and quite another to keep the current system running and making it robust and reliable while trying to make improvements. There are some new hires at NSIDC, including data curators, to make sure data inventory is properly organized and that we are providing the proper levels of service to the data that we have now and will acquire in the future.
It is important to note that both CADIS and ACADIS have been very strong team efforts between the groups at NSIDC and the NCAR groups: Earth Observing Laboratory (EOL); the Cyber Infrastructure Laboratory (CISL); and Unidata, which is within University Corporation for Atmospheric Research (UCAR). The point is that the team working on this project involves 20 people who have lent their expertise and unique perspective to the development of the original system and now continue to work in ACADIS. Developing and keeping a team like this together is very unique. We are really proud of that.
WTA: Tell us about the Data Advisory Committee.
Serreze: The core group of the committee is in place and includes:
- Andrew Slater, NSIDC, land surface processes;
- Dave Bailey, NCAR, global climate modeling;
- Larry Hamilton, representing SEARCH and the social sciences;
- Mary-Louise Timmermans, Yale University, deep-water oceanography;
- Craig Tweedie, University of Texas at El Paso, terrestrial ecology; and
- Carin Ashjian, Woods Hole Oceanographic Institution, ocean ecology.
We still need to broaden the perspective, for example, with representation from the paleoclimate community. The committee must have a wide range of arctic researchers who come with different perspectives on data. We are really interested in knowing what their stories are and their frustrations in dealing with data, finding the data, and using the data. Without getting that information, it's going to be difficult for us to move forward.
Moore: The committee has a group of great people on it and we're very optimistic about their participation to make ACADIS more useful as a community resource. I think it's a work in progress to see what kind of advice and coordination they will be providing and how we address these things in ACADIS.
WTA: What are some of the challenges related to developing ACADIS?
Serreze: The next issue to address is data inventory. Working with the managers at NSF, we need to inventory what data has been collected by arctic researchers, where that data is being stored, and what condition it is in. Basic information like this is needed before we can really get the ACADIS system working. There's a ton of data out there and there are other homes that people have been using for their arctic data. We need to discover where they are so we can at least point to them. One of the goals from NSF is to establish greater access to historical datasets so that investigators doing modeling work aren't limited to a few individual datasets scattered all over the place in different locations.
Moore: Our system needs to be able to protect the datasets of the PI while making sure they are easily accessible in the long run for the whole community. What we will do in the cases where the data needs to be protected is offer the ability to password-protect it from premature use by other groups. The PIs make the decision about what data will be available to the public and when that will happen. PIs should not be fearful about submitting their data to the archive. All datasets are made public eventually as required by the NSF data policy.
WTA: What can the arctic research community do to support ACADIS?
Serreze: Simply get involved in the process. For many arctic investigators, the issue of data management is an afterthought. We have to change the way we're thinking about data—we have to realize that the data always needs to be accessible.
Moore: Talk to us! Tell us what your needs are and how we can help you. Keep us informed if there is something not going right, and let us know when things are going right.
WTA: What is your vision of how ACADIS will enhance support for arctic science?
Moore: We are pretty excited about things ACADIS can offer over its lifetime: long-term stewardship of incredibly rich sources of information on the current condition of the Arctic and how the Arctic is changing; a user interface that allows the data provider to easily put data, metadata, and documentation into the system; and as easily allows users—the young scientist, the graduate student, the person doing synthesis or integration of data—to peruse what datasets are out there, either in our system or others around the world, and access and use it. We hope to offer some visualization tools, and offer the data in common formats that would permit an increased ability to see multiple datasets and multiple data types to help with the interpretation of the changes going on in the Arctic.
Serreze: Interdisciplinary data discovery will enable people to integrate data better than they can today. Arctic science is becoming increasingly interdisciplinary. When we think about the challenges of understanding climate change, it means different disciplines working together. So the data management end of that really needs to follow suit. That's the really big picture.