| 22 Jul 2026
Improving Shortwave Radiative Transfer in the Lake Component of the Common Land Model (CoLM-Lake) with a Spectral Scheme for the Full Snow-Ice-Water Column
Abstract. Shortwave radiation controls lake thermal structure and ice evolution, yet one-dimensional lake models typically use bulk Beer–Lambert attenuation in water and optically opaque snow and ice. Here, treating the full snow-ice-water column as a unified, optically active system, we couple the Snow, Ice, and Aerosol Radiation Model with Adding-Doubling version 5 (SNICAR-ADv5) with the lake component of the Common Land Model (CoLM-Lake) to represent spectral radiative transfer for lake columns. The coupled model accounts for wavelength-dependent absorption and scattering, direct and diffuse pathways, Fresnel interface effects, and radiative interactions among layers. We evaluate the coupled model across nine lakes spanning depth, clarity, climate, and ice cover. Its impacts are physically consistent but lake dependent. During open-water periods, the coupled model shifts absorbed shortwave energy toward near-surface, especially for diffuse radiation, and reduces heating below the mixed layer. This behavior improves temperature profiles and thermocline depths where the original model overheats below the mixed layer and places thermoclines too deep. In turbid lakes, observation-based water extinction coefficients provide first-order improvements, whereas the spectral scheme gives a more interpretable heating profile than single-band Beer–Lambert attenuation. During ice-covered periods, it partitions shortwave energy among snow, ice, and water, enabling representation of internal ice heating, under-ice warming, and basal melt. Limited improvements in some deep or clear lakes indicate that radiative transfer advances must be combined with improved mixing and heat storage. This implementation provides a process-based tool for diagnosing and improving lake thermal, ice, and lake–atmosphere interaction simulations in land, weather, and Earth system models.
The incoming shortwave radiation at the lake surface and its partitioning within the snow–ice–water continuum play important roles in lake freeze–thaw evolution and lake thermal dynamics during the open-water period. However, these processes remain inadequately represented in many current lake models. In this study, the authors treat the entire snow–ice–water column as an integrated optically active system and couple the Snow, Ice, and Aerosol Radiative model (SNICAR-ADv5) with the lake component of the Common Land Model (CoLM-Lake) to simulate spectrally resolved radiative transfer through the vertical lake column. This work represents a substantial improvement in the representation of lake radiative processes and has important implications for improving simulations of lake–atmosphere interactions. The topic fits well within the scope of the GMD journal, and the study is scientifically valuable. Nevertheless, several relatively minor issues should be further clarified or discussed before publication.
1. Figure 13, the authors state earlier in the manuscript that observational data on lake-ice thickness are available for some lakes. However, these observations do not appear to be fully incorporated into the evaluation presented in Figure 13, particularly for Nam Co. Previous studies generally indicate that the maximum ice thickness of Nam Co is approximately 0.5 m, whereas the simulations presented here appear to produce considerably thinner ice, especially in the CoLM-SNICAR experiment, which incorporates the more physically advanced radiative scheme. The authors are encouraged to provide a quantitative comparison between the simulated and observed ice thicknesses where possible and to discuss the potential causes and implications of the apparent underestimation.
2. How were the observed lake freeze-up and break-up dates reported in Table 2 determined? For medium and large lakes, ice formation and disappearance generally exhibit considerable spatial heterogeneity, and the corresponding dates may differ substantially among different parts of the lake. The authors should clarify the data sources, identification criteria or thresholds, spatial representativeness, and temporal resolution used to determine these dates.
3. In many previous lake-modeling studies, particularly one-dimensional or single-column simulations, the vertical mixing coefficient within the water column has often been artificially increased by several times or more to obtain more realistic simulations of lake temperature and mixed-layer depth. The authors are encouraged to discuss why such large adjustments are frequently required. Do they primarily indicate limitations in the original vertical mixing parameterization, or do they compensate for the inability of one-dimensional models to represent three-dimensional circulation, horizontal advection, and lateral mixing in large lakes? The vertical mixing parameters adopted in the present study and their physical justification should also be clarified.
4. Regarding the influence of snow cover on lake freeze–thaw simulations, continuous in situ measurements of snow depth or snow water equivalent over lake ice are admittedly scarce. Nevertheless, remotely sensed albedo and visible imagery could still provide useful constraints on the duration, extent, and spatial variability of snow cover over lake ice. Such information could serve as independent evidence for evaluating model assumptions and interpreting simulation errors. Because winter solar radiation remains strong over the Tibetan Plateau, snow cover can substantially influence lake freezing and thawing by modifying surface albedo, radiation penetration, and the thermal evolution of lake ice. The authors are therefore encouraged to use available remote-sensing information to constrain snow conditions where possible, or to provide a more thorough discussion of the uncertainties associated with the prescribed or simulated snow cover.
5. In the discussion of future research, attention should also be given to horizontal exchanges of energy and mass, in addition to vertical radiative transfer and heat exchange within the snow–ice–water column, particularly for medium and large lakes. Horizontal advection, lake circulation, and lateral heat transport can directly affect the lake thermal structure and may partly explain why enhanced vertical mixing is often required in one-dimensional lake models. It would be useful for the authors to expand the discussion of these processes in the limitations and future research section.