by Theresa Pepin
Vast expanses of Arctic permafrost hold the key to the future as our climate changes. What will befall our world as these great, hulking masses of permafrost thaw? How do we learn enough on such a grand scale to really understand what may happen?
So far, research and simulations have just “touched the surface” of the challenge due to its sheer magnitude. Permafrost degradation and its potentially large carbon-climate feedback cycle are still poorly understood. This JDRD-LDRD collaboration pairs a deeper knowledge of the physical processes involved in ground freezing and thawing on a large scale with massively parallel computation to more closely model the impact of global warming on sensitive Arctic regions.
Ed Perfect has studied the physics of ground freezing and thawing phenomena at different spatial scales since his PhD dissertation work in the early 1980s. He has significant experience with the techniques and mathematics employed in predicting the scale dependency of parameters needed to model freezing and thawing of geomaterials. Lessons learned in a previous JDRD project involving neutron imaging are also proving valuable, enabling both radiography and tomography of thawing permafrost samples to yield better clues to the physical processes involved.
Perfect’s JDRD team member and post-doc Chu-Lin Cheng contributes experience in modeling groundwater flow and surface-water/ground-water interactions, using various numerical packages. The team also includes Ken Christle, a summer intern, who brings an apt set of skills to the research tasks of code and parameter-sensitivity analyses.
LDRD team leader Richard Mills just happens to have completed his undergraduate degree in geology and physics at UT. Today he is a computational scientist in the Computational Earth Sciences Group of the Computer Science and Mathematics Division and in the Earth and Aquatic Sciences Group of the Environmental Sciences Division at ORNL. His team aims to improve the state-of-the-art by integrating detailed models of ground freezing and thawing mechanisms, such as those developed by Perfect’s JDRD research team, into the massively parallel subsurface flow and reactive transport code PFLOTRAN.
The collaboration projects led by Perfect and Mills will also enable comparison of simulation results from different software packages—especially those produced by the older STOMP/MarsFlo code with the massively parallel PFLOTRAN. This research promises to yield particular value in the application of neutron imaging to the task of working up to larger scale data from smaller scale data on the continuum of “bench” to field to region. A target dataset is laboratory data generated from the Next-Generation Ecosystem Experiment Arctic project.
Moving beyond the over-simplification of current land surface models, information-rich surface-to-subsurface thermal, hydrologic, and biogeochemical reaction models will be coupled with comprehensive models for land-surface processes using leadership-class supercomputers. The improved code for field-to-regional-scale simulations will be known as CLM-PFLOTRAN (Community Land Model-PFLOTRAN). For UT and ORNL, the anticipated result will be an exciting and completely new initiative in the strategically core research areas of climate change and supercomputing.
Coupled simulation of hydrologic processes and terrestrial ecosystem and climate feedbacks: Inclusion of soil freezing/thawing and upscaling modules in PFLOTRAN
Ed Perfect, UT Department of Earth and Planetary Sciences
Coupled simulation of surface-subsurface hydrologic processes and ecosystem and climate feedbacks: from Arctic landscapes to the continental US
Richard Mills, ORNL Computer Science and Mathematics Division