Beijing Climate Center Climate System Model 1.1 - m

Book describing model components: http://www.springer.com/us/book/9789811000317

From Section 1.5.1.3 Development of the High-Resolution Land Component Model:

BCC_AVIM2.0 is the newly developed version 2 of the BCC’s Atmosphere– Vegetation Interaction Model, and is suitable for inclusion in BCC_CSM2. Its previous version, BCC_AVIM1.0, was developed on the basis of version 3 of the NCAR’s Community Land Model (CLM3.0) and AVIM2, and includes: (1) A soil moisture and heat transfer module similar to that of CLM3.0, with its underlying surfaces falling into four categories: soil, wetland, lake, and glacier.

No where else in this document is the term "lake" mentioned.  Presumably, the CLM3.0 documentation may provide more details.

From Section 1.2.3 of CLM3.0 Documentation: Surface Data Required

surface data for each land grid cell are listed in Table 1.3 and include the glacier, lake, wetland, and urban portions of the grid cell

Percent lake and wetland were derived from Cogley’s (1991) 1.0º by 1.0º data for perennial freshwater lakes and swamps/marshes. 

CLM3.0 Document Section 9 describes the lake model:

The lake model is from Zeng et al. (2002), which utilized concepts from the lake models of Bonan (1996), Henderson-Sellers (1985, 1986), Hostetler and Bartlein (1990) and the coupled lake-atmosphere model of Hostetler et al. (1993, 1994). All lakes are currently “deep” lakes of 50 m depth. Temperatures are simulated for ten layers with layer thicknesses of 0.1, 1, 2, 3, 4, 5, 7, 7, 10.45, and 10.45 m, and node depths located at the center of each layer (i.e., 0.05, 0.6, 2.1, 4.6, 8.1, 12.6, 18.6, 25.6, 34.325, 44.775 m). Lake surface fluxes closely follow the formulations for non-vegetated surfaces (section 5.2). The lake surface temperature i ∆z i z T g is solved for simultaneously with the surface fluxes. Snow on lakes is based on a bulk approach, not on the multi-layer model described in section 7.2. 

Cogley, J.G. 1991. GGHYDRO – Global Hydrographic Data Release 2.0. Trent Climate Note 91-1, Dept. Geography, Trent University, Peterborough, Ontario. 

From the GGHYDRO site, the lake variable would appear to be FLAK:

The GGHYDRO data are .LIS files and there is a reference to Fortran for reading them. 

possible resources:

Bonan, B.G., 1995, Sensitivity of a GCM simulation to inclusion of inland water surfaces, Journal of Climate, 8, 2691-2704.

Figure 1 is a map based on GGHYDRO but lower resolution

Coe, M.T., 1998, A linked global model of terrestrial hydrologic processes: simulation of modern lakes, rivers and wetlands, Journal of Geophysical Research, 103, 8885-8899.

This figure is from Coe (1998) and shows North American lakes reprojected to a 5'x5' resolution from Cogley [1991].  The Great Lakes are obvious.

In a presentation about the model there is mention that a deep-lake module will be incorporated into BCC_AVIM2 which is part of CMIP6.  This leads me to believe lake dynamics are not simulated in the current version of the model.