Extended Collaborative Support Service (ECSS)
Please find below news stories published by XSEDE and its affiliates regarding the Extended Collaborative Supports Services (ECSS).
Climate Hydro Models Conent
Bridging models and communities
ECSS couples climate, hydro models on Kraken
October 30, 2012
One of XSEDE's primary objectives is making sure that researchers can get the most research bang for their buck on member resources.
And one of the primary tools for achieving this objective is the Extended Collaborative Support Service (ECSS), which pairs members of the XSEDE user community with expert (ECSS) staff at member institutions for an extended period to work together to solve challenging science and engineering problems.
One of the program's more recent success stories is soon to be felt in the climate arena. By allowing a hybrid team of researchers from the hydrological surface and climate communities to access Kraken, a Cray XT5 located at the National Institute for Computational Sciences (NICS), XSEDE is paving the way for novel research approaches that bring together seemingly separate fields for coordinated investigations.
"In the last few years, I had been interacting with the hydrological community and learned about their frameworks," said Cecilia DeLuca, a computational scientist at the National and Oceanic Atmospheric Administration and a principal investigator on the project. "With the emergence of interest in climate impacts I thought that it would be interesting to try to do a two-way coupling between a high-performance atmosphere and hydrological surface model without putting them in the same model ... I wanted to bridge the models and communities."
Essentially, DeLuca wanted to couple the very complex Community Earth Systems Model (CESM), a global climate model that provides state-of-the-art computer simulations of the Earth's past, present, and future climate states, with the hydrological surface model known as the Soil and Water Assessment Tool (SWAT), a river basin scale model developed to quantify the impact of land management practices in large, complex watersheds.
The reason is simple: climate change has an impact on regional hydrology, and those changes will in turn affect climate change. "Without a two-way coupled system you really can't model those feedbacks," said DeLuca.
At first, however, she had her reservations. "I didn't have a good feeling of time and spatial scales, but I had the sense that spatial scales are converging since climate modelers are zeroing in while hydrological models are getting larger," she said. In fact, DeLuca performed a scaling analysis and learned that the CESM/SWAT coupling work would be increasingly feasible as the models evolved. However, more obstacles remained.
The research team, led by Dr. Richard Rood the University of Michigan, tried numerous computing centers before receiving their XSEDE allocation. Time after time, security was the issue, including with Kraken, as remote access isn't permitted on the petaflop-plus powerhouse. Enter XSEDE and the ECSS. By putting their figurative heads together, XSEDE and NICS user support came up with an ingenious solution that allowed Rood's research team access to a leadership-class system.
While the problem might not seem hard on the surface, said NICS's Haihang You, "it's difficult because of security requirements on supercomputers." You coordinates the research team's allocation and communicates between them, NICS, and XSEDE.
DeLuca originally had the idea to use a web service to circumvent the security issues, but didn't know if it was feasible. NICS proved that not only was it feasible, it was actually practical.
Essentially, said You, Indiana University provides a virtual webserver for the research team, which they then connect to via a desktop program. From there an SGI card connects them to Kraken. The project has worked so well that the research team is looking to expand its base and recruit more investigators to refine the models and produce even better results.
Better yet, this approach is extendable to other areas of climate such as fire and drought. What makes this approach different, said DeLuca, is that it enables the hydrology community to use a system its familiar with and develop products while still getting valuable scientific feedbacks. "If you just did the whole thing in CESM, it would hamper the hydro community's efforts," she said.
The success of the collaboration is a testament to the ECSS program and XSEDE's philosophy of working with researchers to ensure that the vast resources at their disposal are being used efficiently. It's also a good omen of things to come; when everyone works together, great things happen.
Gregory Scott Jones
National Institute for Computational Sciences