- Working Groups
At the Winter 2013 LIWG meeting (Boulder, CO, Feb. 14-15), there was an open discussion about strategic and model development priorities for the LIWG during the upcoming year. As part of that discussion, the following goals and priorities were agreed on.
(1) The LIWG needs to release a robust, parallel, efficient, stand-alone dycore (CISM 2.0) that is well documented and well supported for both a range of test cases that stress standard code configurations (e.g., different combinations of boundary conditions, flow regimes, etc.) and model configurations that are useful for science applications (e.g., configurations and initial conditions for Greenland and/or Antarctica in approximate steady-state with pre-industrial climates).
(2) For the May 2013 release of CESM 1.2, we will aim to include the CISM 2.0 code base but only officially support CISM 1.0 physics. All standard CISM 1.0 test cases and several standard science configurations (e.g., 5 km Greenland with spun-up initial conditions and namelist settings) will also be supported. CISM 2.0 physics (higher-order, parallel dynamical core), standard test cases, and possibly some science configurations (including initial conditions) may be included in the release but will not be scientifically validated or officially supported. This may involve segregation of the 2.0 physics from the 1.0 physics within the directory structure so that we can easily include or exclude 2.0 physics from the CESM build. The extent to which CISM 2.0 physics are supported will be dictated by the progress made towards a stand-alone CISM 2.0 release. In addition, we will aim to release the appropriate “hooks” (e.g., software interfaces) to the BISICLES dynamical core in the version of CISM that is released with CESM 1.2. This will not only allow the LIWG to easily continue work on coupling of BISICLES to CISM and CESM but will also serve as a template for other land ice modeling groups (who may eventually wish to couple dynamical cores to CISM and CESM).
(3) The LIWG will aim to place additional emphasis on support and development necessary for coupled ice sheet and climate model applications, which is a potential science niche for CESM and the LIWG (relative to other stand-alone land ice modeling groups and/or climate modeling groups). Similarly, the LIWG will place additional emphasis on supporting the scientific needs of the paleoclimate community within CESM. By necessity, first efforts in these areas may be initially restricted to dynamical cores with “shallow ice approximation” (CISM1) physics. This may require additional development support for the SIA dynamical core (e.g., the ability to run the SIA core in parallel). Early successes using CISM1 physics will be used to guide future work using more expensive and fragile CISM2 physics.
(4) The LIWG will continue to aggressively pursue the still novel area of land ice / ocean model coupling, within both its current (POP and CISM) and next-generation (MPAS-Ocean and MPAS-land ice) model components. This is an area where the LIWG and CESM maintain a slight edge relative to other modeling groups, and the LIWG and CESM should work hard to take advantage of the high-profile science that could be done as a result of this edge. In the short term coupling of the BISICLES dycore in CISM could lead to some high-profile science results.
(5) The LIWG will work to improve CISM documentation and CISM’s web presence in general, including online documentation and tutorials. It was recognized that there is currently a critical support gap for this work. In the future, all CISM developers and users will be encouraged to use automatic documentation tools, wikis, and other similar tools for keeping documentation and online materials accurate, sufficiently detailed, and up to date.
(6) In addition to playing a stronger role in the standard yearly CESM tutorial, in the future, the LIWG may want to consider holding at least one yearly CISM-specific tutorial, in order to build more interest within the science community for using CISM in non-CESM-specific applications (e.g., stand-alone model runs driven by offline forcing; physical process models being developed inside of the stand-alone model).
(1) We will release a robust, parallel, efficient, stand-alone dycore (CISM 2.0) that is well documented and supported for a range of test cases and scientifically useful model configurations.
(2) We will include the CISM 2.0 code base within the CESM 1.1.1 release in May of 2013. While we will only officially support CISM 1.0 physics, we will aim to included some support for CISM 2.0, through test cases and possibly through a limited number of model configurations for science. We will also aim to include the software interfaces necessary for coupling the BISICLES dynamical core to CISM 2.0.
(3) We will place additional emphasis on support and development necessary for coupled ice sheet and climate model applications, including the scientific needs of the paleoclimate community within CESM. First efforts in these areas may be restricted to CISM 1.0 physics.
(4) We will continue to aggressively pursue development and testing necessary for land ice / ocean model coupling, within both current and next-generation model components.
(5) We will focus on improving CISM documentation and CISM’s overall web presence, including online documentation and tutorials. We will also aim to play a stronger role in CESM’s yearly tutorial and, as new dynamical cores and model features come online, consider holding semi-regular tutorials that focus more on the needs of stand-alone model users.