Spring 2018 – Guest Frequently Asked Questions with Dr. Peter Brewer
– JUNE 19, 2018
(From Spring 2018 Newsletter) Dr. Peter Brewer, GoMRI Research Board Member and co-chair of the GoMRI Data Management Committee, answered a few questions about the Gulf of Mexico Research Initiative Information and Data Cooperative (GRIIDC) and the legacy of GoMRI and GRIIDC’s data policies.
Question: What is the GoMRI Data Management Subcommittee, and what has been your role as co-chair?
Answer: The GoMRI Research Board (RB) provides leadership and guidance to the overall GoMRI program. Initially the RB tended to act as a Committee of the Whole, but it was quickly realized that more specific assignments for RB members were needed. I happened to be tasked, along with the other Data Management Committee members David Halpern, Ken Halanych, Margaret Leinen, and Burt Singer, with oversight over our data management activities. We each brought different skills and backgrounds to this task. I did have some large program management experience both scientifically and administratively so this seemed to be a fair assignment. Initially this was confusing since everyone was newly acquainted, and a working basis had yet to be established. One important, very early event was a visit by GRIIDC Director Jim Gibeaut and colleagues from their home base at the Harte Institute for Gulf of Mexico Studies at Texas A&M University-Corpus Christi to my own lab in Moss Landing, California. This allowed ease of introduction and a common understanding of goals to be established on a personal basis. As Data Management Committee members, we have to work with both RB colleagues and the staff of the GRIIDC team. It has been a pleasure to do so.
Question: What are GoMRI’s data sharing policies? What do you think are the most important aspects of those policies and why?
Answer: GoMRI’s data sharing policies are not on their face remarkable. What is remarkable is that these were not already widely embedded in our ocean science community. The special funding and unique oversight provided by generous support from BP through the Gulf of Mexico Alliance (GOMA) has made possible enforcement of eminently sensible rules by the RB. When you publish a scientific paper, you are also implicitly publishing the actual data behind your arguments so that the truth can be directly observed. Unfortunately, the explicit implementation of this simple rule had become widely flouted; data sets had become large, and the old style printed list of numbers in a publication no longer works. And scientists had become accustomed to illusions of “MY data” and the excuse that they were just too busy to bother with requests for copies. GoMRI changed all that, and the timing just preceded the national awakening to the problem as identified in the May 9, 2013 Executive Order establishing a national open data policy. All those who are awarded research funding from GoMRI are informed in their award letter that they are bound by full data submission requirements and that the goal is to have the data supporting any and all publications openly available on the GRIIDC website the very day of publication. This allows scientific claims to be validated against the numbers and gives readers a chance to work with the data to test pet theories, investigate alternate explanations, or educate students by setting these data sets as class problems for study.
That this has sadly not been the norm in science was illuminated when, in frustration in comparing the success of GoMRI-GRIIDC to the attitudes of others, I wrote a piece for the American Geophysical Union’s (AGU) Eos in a fit of pique. See: https://eos.org/articles/read-them-again-eoss-most-viewed-stories-of-2017 (Article #8). That article has now been viewed some 10,000 times!
What has been a pleasure is to see how happily this has all worked out with huge credit to the GRIIDC team for their gentle but firm coaching of principal investigators (PIs) and the careful way the path to success has been laid. It is only as a last resort that we inform researchers that we will cut off their funding if they do not comply; thankfully, this rarely, if ever, happens. We have made compliance relatively painless, and it works.
Question: How has GoMRI’s data model, as executed by GRIIDC, advanced science and expectations of scientists?
Answer: The expectations of scientists have changed considerably, and for the better. Data archiving beyond one’s personal data base is now the norm. So too has the expectation that you are not on your own here and that questions can be asked, advice given, and that common standards can be made to work very well. The GRIIDC team deserves huge credit here for immediately seeing their role as coach, teacher, and colleague – not as the enforcer. Initially there was a blizzard of questions, often messy: what about model output “data,” what about instrument settings, at what level do we use “raw” data versus “processed” data, what about different versions of a data set, etc. All of these have had to be negotiated carefully and with a useful result. The very early insistence on metadata – the critical context within which measurements are made – has paid off handsomely. It is often the first point of entry into the system, and GRIIDC has handled this well.
Question: What will be the legacy of GRIIDC after the GoMRI program ends in 2020?
Answer: The Research Board is already planning for a form of continuation of GRIIDC in order to preserve the data legacy for several years (through 2030), although in a reduced role. One hopes that the lessons learned will endure, and that we will no longer have lost data when a PI retires or gets a new job offer and must walk away from the lab where the information was created. One hopes that a fully fleshed out picture of the Gulf will emerge as the GoMRI data are combined with other data resources, such as those from the Natural Resource Damage Assessment (NRDA) activities. This will require a merging of the GoMRI and federal data activities, and good people are working hard on that – although in the face of great organizational difficulties. One hopes that the remarkable results obtained with GoMRI funding can speak loud and clear to the better prediction of the consequences of another oil spill somewhere else in the world and sometime in the future. That means extracting the fundamentals from the observations – not always an easy task. One hopes that the interdisciplinary lessons learned endure: the mating of the molecular interfacial forces between oil and water with the large-scale physics of a plume or surface film, the ability to observe and predict the activities of microbes as they make chemical calculations on the energy available from a rapidly changing “food” supply, the real impact of added dispersants on the fate of the oil and on the organisms. All of these are addressed in some way in GoMRI research – and the details are recorded in the GRIIDC archives. This is the resource that problem solvers of the future must turn to.