Quantifying the skill of CMIP5 models in simulating seasonal albedo and snow cover evolution

Abstract

Effectively modeling the influence of terrestrial snow on climate in general circulation models is limited by imperfect knowledge and parameterization of arctic and subarctic climate processes and a lack of reliable observations for model evaluation and improvement. This study uses a number of satellite-derived data sets to evaluate how well the current generation of climate models from the Fifth Coupled Model Intercomparison Project (CMIP5) simulate the seasonal cycle of climatological snow cover fraction (SCF) and surface albedo over the Northern Hemisphere snow season (September-June). Using a variety of metrics, the CMIP5 models are found to simulate SCF evolution better than that of albedo. The seasonal cycle of SCF is well reproduced despite substantial biases in simulated surface albedo of snow-covered land (α sfc_snow ), which affect both the magnitude and timing of the seasonal peak in α sfc_snow during the fall snow accumulation period, and the springtime snow ablation period. Insolation weighting demonstrates that the biases in α sfc_snow during spring are of greater importance for the surface energy budget. Albedo biases are largest across the boreal forest, where the simulated seasonal cycle of albedo is biased high in 15/16 CMIP5 models. This bias is explained primarily by unrealistic treatment of vegetation masking and subsequent overestimation (more than 50% in some cases) of peak α sfc_snow rather than by biases in SCF. While seemingly straightforward corrections to peak α sfc_snow could yield significant improvements to simulated snow albedo feedback, changes in α sfc_snow could potentially introduce biases in other important model variables such as surface temperature.

Why Do Models Produce Spread in Snow Albedo Feedback?

“…Simulated vegetation bias is linked to the surface land cover data set used to derive vegetation cover and/or issues stemming from implementation of dynamic vegetation schemes, which 11 of 27 CMIP5 models feature (Table 1). In contrast, the vegetation scheme within the Model for Interdisciplinary Research on Climate (MIROC) models generates a very low leaf area index over the NH extratropics (Wang et al, 2016) exposing more bare ground and inflating α sc (Loranty et al, 2014;Thackeray et al, 2015). In contrast, the vegetation scheme within the Model for Interdisciplinary Research on Climate (MIROC) models generates a very low leaf area index over the NH extratropics (Wang et al, 2016) exposing more bare ground and inflating α sc (Loranty et al, 2014;Thackeray et al, 2015).…”

Vegetation masking effect on future warming and snow albedo feedback in a boreal forest region of northern Eurasia according to MIROC‐ESM

“…However, because of the insufficient observed snowfall and snowmelt data, the cause is unclear. Furthermore, the surface albedo over the northern terrestrial regions in MIROC models is higher than that in other models [ Loranty et al ., ; Thackeray et al ., ]. This higher surface albedo effect might enhance these biases.…”

Inferring Surface Albedo Prediction Error Linked to Forest Structure at High Latitudes

“…Sellers, ). While recent investigations have typically focused on the prediction error arising from differences in model parameterizations of albedo (Bartlett & Verseghy, ; Boisier et al, ; Bright et al, ; Thackeray et al, , ), model representations of snow physical attributes and/or spatial coverage (Boisier et al, ; Y. Li et al, ; Thackeray et al, ), or from differences in the vegetation mapping (Boisier et al, ), few studies have evaluated the prediction error attributable to differences in model representations of vegetation structure. Recent attribution analyses elsewhere have suggested that errors in simulated vegetation structure could be an important source of the persistent albedo prediction error seen in the current generation of climate models (Y. Li et al, ; Wang et al, ).…”