Variation in snow albedo feedback (SAF) among CMIP5 climate models has been shown to explain much of the variation in projected 21st Century warming over Northern Hemisphere land. Prior studies using observations and models have demonstrated both considerable spread in the albedo, and a weak bias in the simulated strength of SAF, over snow-covered boreal forests. Boreal evergreen needleleaf forests are capable of intercepting snowfall throughout the snow season, which has a significant impact on seasonal albedo. Two satellite data products and tower-based observations of albedo are compared with simulations from multiple configurations of the Community Climate System Model (CCSM4) to investigate the causes of weak simulated SAF over the boreal forest. The largest bias occurs in April-May when simulated SAF is one-half the strength of SAF in observations. This is traced to two canopy snow parameterizations in the land model. First, there is no mechanism for the dynamic removal of snow from the canopy when temperatures are below freezing, which results in albedo values in midwinter that are biased high. Second, when temperatures do rise above freezing, all snow on the canopy is melted instantaneously, which results in an unrealistically early transition from a snow-covered to a snow-free canopy. These processes combine to produce large differences between simulated and observed monthly albedo, and are the sources of the weak bias in SAF. This analysis highlights the importance of canopy snow parameterizations for simulating the hemispheric scale climate response to surface albedo perturbations. A number of new experiments are described as recommendations for future work.

Land carbon sensitivity to atmospheric CO2 concentration (bL) and climate warming (gL) is a crucial part of carbon-climate feedbacks that affect the magnitude of future warming. Although these sensitivities can be estimated by earth system models, their dependence on model representation of land carbon dynamics and the inherent model assumptions has rarely been investigated. Using the widely used Community Land Model version 4 as an example, we examine how bL and gL vary with prescribed versus dynamic vegetation covers. Both sensitivities are found to be larger with dynamic compared to prescribed vegetation on decadal timescale in the late twentieth century, with a more robust difference in gL. The latter is a result of dynamic vegetation model deficiencies in representing the competitions between deciduous versus evergreen trees and tree versus grass over the tropics and subtropics. The biased vegetation cover changes the regional characteristics of carbon-nitrogen cycles such that plant productivity responds less strongly to the enhancement of nitrogen mineralization with warming, so more carbon is lost to the atmosphere with rising temperature. The result calls for systematic evaluations of land carbon sensitivities with varying assumptions for land cover representations to help prioritize development effort and constrain uncertainties in carbon-climate feedbacks.