The model is designed to run in three different configurations:

1. Stand-alone executable code as part of the Community Climate System Model (CCSM).

2. A subroutine call within the Community Atmosphere Model (CAM) in which CAM/CLM represent single executable code.

3. Stand-alone executable code in which the model is forced with atmospheric datasets. In this mode, the model runs on a spatial grid that can range from one point to global.

The model is designed to integrate all land processes into a single model. This is consistent with the coupling philosophy of the Community Climate System Model, which recognizes atmosphere, land, ocean, and sea ice as the core component models communicating among each other through a coupler. Initially, the land model only considered biogeophysics (i..e., surface fluxes of energy, moisture, momentum) and hydrology. As the model expands to include river routing, biogeochemistry (e.g., dust, volatile organic compounds), vegetation dynamics, and the carbon cycle, the issue of how these new land processes are brought into the model arises. Each of these components exists as a separate scientific field of study with their own models and modeling community. One possibility is to hang each new component off the coupler (see figure). However, this maximizes message passing through the coupler because the component models share many common variables. Moreover, the components share many common processes and surface datasets. The models also have to be initialized and write the necessary restart and history files. This is an awkward way to add new land parameterizations to CCSM.

The approach adopted for the Community Land Model adds new component processes as modules or subroutines within the CLM (see figure). This minimizes message passing to and from the coupler because most variables are internal to the land model. It provides standard code to initialize all components, write history files, and read/write restart files. It is much easier in this configuration to add new parameterizations to the model.