Quoted from: https://escomp.github.io/cism-docs/cism-in-cesm/versions/release-cesm2.0/html/introduction.html 

Starting in 2006, researchers in the Climate, Ocean and Sea Ice Modeling (COSIM) group at Los Alamos National Laboratory (LANL) worked with scientists at the National Center for Atmospheric Research (NCAR) to incorporate an ice sheet model in the CCSM/CESM framework. This work was funded primarily by the Department of Energy (DOE) Scientific Discovery through Advanced Computing (SciDAC) program, with additional support from the National Science Foundation (NSF). The Glimmer ice sheet model (Rutt et al. 2009), developed by Tony Payne and colleagues at the University of Bristol, was chosen for coupling. Although Glimmer’s dynamical core was relatively basic, a higher-order dynamics scheme was under development. In addition, the model was well structured and well documented, with an interface (Glint) to enable coupling to ESMs.

In 2009, the U.K. researchers who designed Glimmer joined efforts with U.S. scientists who were developing a Community Ice Sheet Model (CISM), and the model was renamed Glimmer-CISM. Model development was overseen by a six-member steering committee including Magnus Hagdorn (U. Edinburgh), Jesse Johnson (U. Montana), William Lipscomb (LANL), Tony Payne (U. Bristol), Stephen Price (LANL), and Ian Rutt (U. Swansea).

Glimmer had a positive-degree-day (PDD) scheme, which uses empirical formulas to relate surface temperatures to summer melting. PDDs schemes, however, are not ideal for climate change modeling, because empirical relationships that are valid for present-day climate may not hold in the future. Instead, a surface-mass-balance scheme for ice sheets was developed for the Community Land Model (CLM). This scheme computes the SMB in each of ~10 elevation classes per grid cell in glaciated regions. The SMB is passed via the coupler to the ice sheet component, where it is averaged, downscaled, and used to force the dynamic ice sheet model at the upper surface. (See Section 4 for details.)

In 2010, CESM1.0 was released with an initial implementation of ice sheets (Lipscomb et al. 2013). The dynamic ice sheet model was a close approximation of Glimmer-CISM version 1.6, a serial code with shallow-ice dynamics. Optionally, the SMB was computed by CLM in multiple elevation classes for glaciated regions.

Subsequent ice-sheet model development targeted for CESM was led by DOE-funded researchers at LANL, Oak Ridge National Laboratory (ORNL), and Sandia National Laboratories (SNL). Much of this work was done under the Ice Sheet Initiative for CLimate ExtremeS (ISICLES) project, followed by the Predicting Ice Sheet and Climate Evolution at Extreme Scales (PISCEES) project, culminating in the 2014 release of CISM 2.0 (Price et al. 2014). CISM2.0 was a major advance over Glimmer, with support for parallel simulations using higher-order dynamics, as well as a suite of test cases and links to third-party solver libraries.

Subsequently, CISM development shifted to NCAR, with primary support from the National Science Foundation. Recent work has focused on making the model more practical and robust for century-to-millenial scale Greenland simulations. CISM2.1, released concurrently with CESM2.0 in 2018, includes an efficient depth-integrated velocity solver and more realistic options for basal sliding and iceberg calving, among other innovations. CISM participated in the initMIP-Greenland project on ice sheet initialization (Goelzer et al. 2018), and a developmental model version was used for the follow-up initMIP-Antarctica experiments.

Meanwhile, other parts of CESM have evolved to support land-ice science. In CESM2.0 the surface mass balance of the Greenland and Antarctic ice sheets is computed by default in CLM using multiple elevation classes, in simulations with or without dynamic ice sheets. Also, CESM now supports interactive coupling of CISM with CLM, allowing the land topography and surface types to evolve as the ice sheet advances and retreats. The surface melt climate of both ice sheets has improved with the inclusion of a deep firn model that allows for meltwater infiltration and refreezing, as well as realistic firn densification rates. Surface winds over ice sheets are more accurate with a new drag parameterization, and a bias in high-latitude longwave cloud forcing is much reduced.