Impact of absorbing aerosol deposition on snow albedo reduction over the southern Tibetan plateau based on satellite observations

We investigate the snow albedo variation in spring over the southern Tibetan Plateau induced by the deposition of light-absorbing aerosols using remote sensing data from moderate resolution imaging spectroradiometer (MODIS) aboard Terra satellite during 2001–2012. We have selected pixels with 100 % snow cover for the entire period in March and April to avoid albedo contamination by other types of land surfaces. A model simulation using GEOS-Chem shows that aerosol optical depth (AOD) is a good indicator for black carbon and dust deposition on snow over the southern Tibetan Plateau. The monthly means of satellite-retrieved land surface temperature (LST) and AOD over 100 % snow-covered pixels during the 12 years are used in multiple linear regression analysis to derive the empirical relationship between snow albedo and these variables. Along with the LST effect, AOD is shown to be an important factor contributing to snow albedo reduction. We illustrate through statistical analysis that a 1-K increase in LST and a 0.1 increase in AOD indicate decreases in snow albedo by 0.75 and 2.1 % in the southern Tibetan Plateau, corresponding to local shortwave radiative forcing of 1.5 and 4.2 W m⁻², respectively.

     

Albedo of coastal landfast sea ice in Prydz Bay,Antarctica: Observations and parameterization

The snow/sea-ice albedo was measured over coastal landfast sea ice in Prydz Bay, East Antarctica(off Zhongshan Station)during the austral spring and summer of 2010 and 2011. The variation of the observed albedo was a combination of a gradual seasonal transition from spring to summer and abrupt changes resulting from synoptic events, including snowfall, blowing snow, and overcast skies. The measured albedo ranged from 0.94 over thick fresh snow to 0.36 over melting sea ice. It was found that snow thickness was the most important factor influencing the albedo variation, while synoptic events and overcast skies could increase the albedo by about 0.18 and 0.06, respectively. The in-situ measured albedo and related physical parameters(e.g., snow thickness, ice thickness, surface temperature, and air temperature) were then used to evaluate four different snow/ice albedo parameterizations used in a variety of climate models. The parameterized albedos showed substantial discrepancies compared to the observed albedo, particularly during the summer melt period, even though more complex parameterizations yielded more realistic variations than simple ones. A modified parameterization was developed,which further considered synoptic events, cloud cover, and the local landfast sea-ice surface characteristics. The resulting parameterized albedo showed very good agreement with the observed albedo.     

Use of atmospheric radiation measurement program data from Barrow, Alaska, for evaluation and development of snow-albedo parameterizations

Summary Snow albedo is determined from the ratio of out-going to incoming solar radiation using three years of broadband shortwave radiometer data obtained from the Barrow, Alaska, Atmospheric Radiation Measurement (ARM) site. These data are used for the evaluation of various types of snow-albedo parameterizations applied in numerical weather prediction or climate models. These snow-albedo parameterizations are based on environmental conditions (e.g., air or snow temperature), snow related characteristics (e.g., snow depth, snow age), or combinations of both. The ARM data proved to be well suited for snow-albedo evaluation purposes for a low-precipitation tundra environment. The evaluation confirms that snow-age dependent parameterizations of snow albedo work well during snowmelt, while parameterizations considering meteorological conditions often perform better during snow accumulation. Current difficulties in parameterizing snow albedo occur for long episodes of snow-event free conditions and episodes with a high frequency of snow events or strong snowfall. In a further step, the first two years of the ARM albedo dataset is used to develop a snow-albedo parameterization, and the third year’s data serves for its evaluation. This parameterization considers snow depth, wind speed, and air temperature which are found to be significant parameters for snow-albedo modeling under various conditions. Comparison of all evaluated snow-albedo parameterizations with this new parameterization shows improved snow-albedo prediction. Correspondence: Nicole M?lders, Geophysical Institute and College of Natural Science and Mathematics, University of Alaska Fairbanks, 903 Koyukuk Drive, P.O. Box 757320, Fairbanks, AK 99775-7320, USA     

The albedo of snow for partially cloudy skies

The albedo of snow for different cloudiness conditions is an important parameter in the Earth's radiation budget analysis and in the study of snowpack's thermal conditions. In this study an efficient approximate method is derived to calculate the incident spectral solar flux and snow-cover albedo in terms of different atmospheric, cloud, and snow parameters. The global flux under partially cloudy skies is expressed in terms of the clear sky flux and a coefficient which models the effect of scattering and absorption by cloud patches and multiple reflections between the cloud base and snowcover. The direct and the diffuse components of the clear sky flux are obtained using the spectral flux outside the atmosphere and the spectral transmission coefficients for absorption and scattering by molecules and aerosols.The spectral snow reflectance model considers both specular surface reflection and volumetric multiple scattering. The surface reflection is calculated by using a crystal-shape-dependent bidirectional reflectance distribution function; the volumetric multiple scattering is calculated by using a crystal-size-dependent approximate solution in the radiative transfer equation. The input parameters to the model are atmospheric precipitable water, ozone content, turbidity, cloud optical thickness, the size and shape of ice crystals of snow and surface pressure. The model yields spectral and integrated solar flux and snow reflectance as a function of solar elevation and fractional cloudcover.The model is illustrated using representative parameters for the Antarctic coastal regions. The albedo for a clear sky depends inversely on the solar elevation. At high elevations the albedo depends primarily upon the grain size; at low elevation the albedo depends on grain size and shape. The gradient of the albedo-elevation curve increases as the grains become larger and faceted. The albedo for a densely overcast sky is a few percent higher than the clear-sky albedo at high elevations. A simple relationship between grain size and the overcast albedo is obtained. For a set of grain size and shape, the albedo as a function of solar elevation and fractional cloud cover is tabulated.     

Improved Parameterization of Snow Albedo in WRF + Noah: Methodology Based on a Severe Snow Event on the Tibetan Plateau

Snowfall and the subsequent evolution of the snowpack have a large effect on the surface energy balance and water cycle of the Tibetan Plateau (TP). The effects of snow cover can be represented by the WRF coupled with a land surface scheme. The widely used Noah scheme is computationally efficient, but its poor representation of albedo needs considerable improvement. In this study, an improved albedo scheme is developed using a satellite-retrieved albedo that takes snow depth and age into account. Numerical experiments were then conducted to simulate a severe snow event in March 2017. The performance of the coupled WRF/Noah model, which implemented the improved albedo scheme, is compared against the model’s performance using the default Noah albedo scheme and against the coupled WRF/CLM that applied CLM albedo scheme. When the improved albedo scheme is implemented, the albedo overestimation in the southeastern TP is reduced, reducing the RMSE of the air temperature by 0.7°C. The improved albedo scheme also attains the highest correlation between the satellite-derived and the model-estimated albedo, which provides for a realistic representation of both the snow water equivalent (SWE) spatial distribution in the heavy snowbelt (SWE > 6 mm) and the maximum SWE in the eastern TP. The underestimated albedo in the coupled WRF/CLM leads to underestimating the regional maximum SWE and a consequent failure to estimate SWE in the heavy snowbelt accurately. Our study demonstrates the feasibility of improving the Noah albedo scheme and provides a theoretical reference for researchers aiming to improve albedo schemes further.     

A new snow cover fraction parametrization for the ECHAM4 GCM

 Snow cover fraction (SCF) has a significant influence on the surface albedo and thus on the radiation balance and surface climate. Long-term three dimensional simulations with general circulation models (GCMs) show that the SCF greatly affects the climate in the Northern Hemisphere. By means of both ground observations and remotely sensed data, several deficiencies in the SCF simulated by the current ECHAM4 GCM were identified: over mountainous areas a substantial overestimation in the SCF was found whereas flat areas showed a distinctly underestimated SCF. This work proposes a new parametrization of the SCF for use in GCMs. Evaluations illustrate that it is beneficial to distinguish between the following three terrains: (1) flat, non-forested areas, (2) mountainous regions and (3) forests. The modified SCF parametrization for flat, non-forested areas was derived by using global datasets of ground-based snow depth and remote sensing observations of snow cover data. A 3-dimensional ECHAM4 simulation showed that this modification raises the SCF by up to approximately 20%, mainly in areas with a relatively thin snow cover. The comparison between remotely sensed and simulated mean monthly surface albedo revealed a significant overestimation of the surface albedo in snow-covered mountainous areas. An extension of the current SCF parametrization in ECHAM4 to take into account mountain effects, based on the French climate model Arpège, yielded a close agreement with satellite-derived surface albedo. The adoption of the submodel for snow albedo, as used in the Canadian Land Surface Scheme (CLASS), combined with a newly developed simple snow interception model, demonstrated the ability to capture the main physical processes of snow-covered canopies, including the albedo. The validation of the new parametrization with Boreal Ecosystem-Atmosphere Study (BOREAS) field data showed that the modification is appropriate to capture the main features of the albedo over snow-covered forests during and after heavy snowfall events. Furthermore, the proposed modification has a beneficial impact on the delayed snow melt in spring, a well-known problem in many current GCMs: The simulated surface albedo over the boreal forests decreases by approximately 0.1 during winter and spring, which is in better agreement with ground-based observations. This induces a significant rise in the surface temperature over extended parts of Eurasia and North America in late spring, which subsequently yields a faster snowmelt and an accelerated retreat of the snow line. Received: 28 April 2000 / Accepted: 18 December 2000     

利用NOAA—AVHRR资料估算积雪参量

本文介绍了利用气象卫星AVHRR资料对青海省青南高原看牧区冬春（10月－4月）积雪进行监测的原理，提出了利用AVHRR资料估算积雪深度和面积的方法，并对其精度进行了检验。     

Simulation of canopy radiation transfer and surface albedo in the EALCO model

A new canopy radiation transfer and surface albedo scheme is developed as part of the land surface model EALCO (Ecological Assimilation of Land and Climate Observations). The model uses a gap probability-based successive orders of scattering approach that explicitly includes the heterogeneities of stands and crown elements and the radiation multiple scattering. The model uses the optical parameters of ecosystem elements and physically represents ecosystem processes in surface albedo dynamics. Model tests using measurements from a boreal deciduous forest ecosystem show that the model well reproduced the observed diurnal and seasonal albedo dynamics under different weather and ecosystem conditions. The annual mean absolute errors between modeled and measured daily albedo and reflected radiation are 0.01 and 1.33 W m⁻², respectively. The model results provide a quantitative assessment of the impacts of plant shading and sky conditions on surface albedo observed in high-latitude ecosystems. The contribution of ground snow to surface albedo in winter was found to be less than 0.1 even though the canopy is leafless during this time. The interception of snow by the leafless canopy can increase the surface albedo by 0.1–0.15. The model results show that the spectral properties of albedo have large seasonal variations. In summer, the near infrared component is substantially larger than visible, and surface albedo is less sensitive to sky conditions. In winter, the visible band component is markedly increased and can exceed the near infrared proportion under cloudy conditions or when snow exists on the canopy. The spectral properties of albedo are also found to have large diurnal variations under the clear-sky conditions in winter.     

Improvement of a Snow Albedo Parameterization in the Snow–Atmosphere–Soil Transfer Model:Evaluation of Impacts of Aerosol on Seasonal Snow Cover

The presence of light-absorbing aerosols(LAA) in snow profoundly influence the surface energy balance and water budget.However,most snow-process schemes in land-surface and climate models currently do not take this into consideration.To better represent the snow process and to evaluate the impacts of LAA on snow,this study presents an improved snow albedo parameterization in the Snow–Atmosphere–Soil Transfer(SAST) model,which includes the impacts of LAA on snow.Specifically,the Snow,Ice and Aerosol Radiation(SNICAR) model is incorporated into the SAST model with an LAA mass stratigraphy scheme.The new coupled model is validated against in-situ measurements at the Swamp Angel Study Plot(SASP),Colorado,USA.Results show that the snow albedo and snow depth are better reproduced than those in the original SAST,particularly during the period of snow ablation.Furthermore,the impacts of LAA on snow are estimated in the coupled model through case comparisons of the snowpack,with or without LAA.The LAA particles directly absorb extra solar radiation,which accelerates the growth rate of the snow grain size.Meanwhile,these larger snow particles favor more radiative absorption.The average total radiative forcing of the LAA at the SASP is 47.5Wm~(-2).This extra radiative absorption enhances the snowmelt rate.As a result,the peak runoff time and "snow all gone" day have shifted 18 and 19.5 days earlier,respectively,which could further impose substantial impacts on the hydrologic cycle and atmospheric processes.     

Empirical models of monthly and annual albedo in managed boreal forests of interior Norway

An 11-year remotely sensed surface albedo dataset coupled with historical meteorological and stand-level forest management data for a variety of stands in Norway’s most productive logging region is used to develop regression models describing temporal changes in forest albedo following clear-cut harvest disturbance events. Datasets are grouped by dominant tree species, and two alternate multiple regression models are developed and tested following a potential-modifier approach. This result in models with statistically significant parameters (p?<?0.05) that explain a large proportion of the observed variation, requiring a single canopy modifier predictor coupled with either monthly or annual mean air temperature as a predictor of a stand’s potential albedo. Models based on annual mean temperature predict annual albedo with errors (RMSE) in the range of 0.025–0.027, while models based on monthly mean temperature predict monthly albedo with errors ranging between of 0.057–0.065 depending on the dominant tree species. While both models have the potential to be transferable to other boreal regions with similar forest management regimes, further validation efforts are required. As active management of boreal forests is increasingly seen as a means to mitigate climate change, the presented models can be used with routine forest inventory and meteorological data to predict albedo evolution in managed forests throughout the region, which, together with carbon cycle modeling, can lead to more holistic climate impact assessments of alternative forest harvest scenarios and forest product systems.     

青海省甘德两次降雪过程的微气象特征分析

利用青海省甘德两次降雪过程的微气象观测数据,探讨了两场降雪过程雪深、雪密度、雪中含冰量、雪中含水量和雪面温度的变化情况,分析了地表反照率与雪密度、雪中含冰量及雪中含水量的关系,结合降雪过程近地面温、湿、风廓线特征分析了积雪对近地面温、湿、风梯度的影响。结果表明:积雪覆盖会导致地表反照率显著增加,降雪过后正午时地表反照率可高达0.8～0.9。随着积雪的消融,地表反照率逐渐减小;积雪反照率与雪密度和雪中含冰量呈正相关,与雪中含水量呈负相关;地表积雪覆盖会导致近地面温度梯度绝对值减小,相对湿度梯度绝对值在凌晨减小、午后增大,地表积雪覆盖对近地面风速梯度变化并无特定的影响。     

Parametrization schemes of incident radiation in the North Water polynya

Abstract

An evaluation of the Canadian Land Surface Scheme (CLASS) 3.1 snow cover simulations at four sites included in the Snow Model Intercomparison Project (SnowMIP) revealed that CLASS was able to provide realistic representations of snow cover accumulation, melt and physical properties over a range of snow cover climates. The modified snow aging parametrization in CLASS 3.1 provided improved simulations of snowpack density which resulted in a marked reduction in the root‐mean‐square (rms) error for daily snow depth, and slight improvements in snow surface temperature. CLASS 3.1 still exhibited a tendency to overestimate snow cover duration which is attributed to the way shallow snow ablation is treated. CLASS provided generally realistic simulations of daily and seasonal variation in snow albedo although cold snow albedo was underpredicted by 0.10 to 0.15 at a site with a deep (> 2 m) cold snowpack. CLASS also exhibited a tendency to overpredict late spring snow albedo which was reduced by the addition of a snow layer subroutine that kept track of snow albedo by precipitation event. CLASS had a noticeable cold bias averaging 3°–4°C at two mountain sites included in the comparison. The bias was closely linked to atmospheric stability and could exceed 10°C under conditions of strong radiative cooling and low wind speeds. The CLASS energy deficit under these conditions was determined to be ～20–40 W m^?2 and was mostly accounted for by introducing a windless exchange coefficient into the calculation of sensible heat fluxes following the approach used in a number of other physical snowpack models. CLASS provided realistic simulations of daily snowmelt runoff with the exception of the Weissfluhjoch site which was characterized by a deep cold snowpack. A preliminary assessment of snow water equivalent (SWE) rms error for the 23 models participating in SnowMIP showed that CLASS was one of the better single layer snow models included in the comparison. CLASS performance was comparable to the multi‐layer CROCUS snowpack model in the evaluations carried out in this study.     

有关雪盖模型内部及界面过程的参数化方案的敏感试验

In order to develop a seasonal snow model of land surface process as accurately as possible for climatic study, it is necessary to fully understand the effects of important snow internal processes and interaction with air and to get an insight into influence of several relevant parameterization schemes with parameters＇ uncertainty to some degree. Using the snow model (SAST) developed by first author and other one and some useful field observation data, this paper has conducted a series of sensitivity studies on the parameterization schemes. They are relative to compaction process, snow thermal conduction, methodology of layering snow pack and to key parameters such as snow albedo, water holding capacity. Then, based on the results from the sensitivity studies, some useful conclusions for snow cover model improvement are ob tained from the analysis of the results.     

Quantifying uncertainties in projections of extremes��a perturbed land surface parameter experiment

Uncertainties in the climate response to a doubling of atmospheric CO₂ concentrations are quantified in a perturbed land surface parameter experiment. The ensemble of 108 members is constructed by systematically perturbing five poorly constrained land surface parameters of global climate model individually and in all possible combinations. The land surface parameters induce small uncertainties at global scale, substantial uncertainties at regional and seasonal scale and very large uncertainties in the tails of the distribution, the climate extremes. Climate sensitivity varies across the ensemble mainly due to the perturbation of the snow albedo parameterization, which controls the snow albedo feedback strength. The uncertainty range in the global response is small relative to perturbed physics experiments focusing on atmospheric parameters. However, land surface parameters are revealed to control the response not only of the mean but also of the variability of temperature. Major uncertainties are identified in the response of climate extremes to a doubling of CO₂. During winter the response both of temperature mean and daily variability relates to fractional snow cover. Cold extremes over high latitudes warm disproportionately in ensemble members with strong snow albedo feedback and large snow cover reduction. Reduced snow cover leads to more winter warming and stronger variability decrease. As a result uncertainties in mean and variability response line up, with some members showing weak and others very strong warming of the cold tail of the distribution, depending on the snow albedo parametrization. The uncertainty across the ensemble regionally exceeds the CMIP3 multi-model range. Regarding summer hot extremes, the uncertainties are larger than for mean summer warming but smaller than in multi-model experiments. The summer precipitation response to a doubling of CO₂ is not robust over many regions. Land surface parameter perturbations and natural variability alter the sign of the response even over subtropical regions.     

Evaluation of a high-resolution regional climate simulation over Greenland

A simulation of the 1991 summer has been performed over south Greenland with a coupled atmosphere–snow regional climate model (RCM) forced by the ECMWF re-analysis. The simulation is evaluated with in-situ coastal and ice-sheet atmospheric and glaciological observations. Modelled air temperature, specific humidity, wind speed and radiative fluxes are in good agreement with the available observations, although uncertainties in the radiative transfer scheme need further investigation to improve the model’s performance. In the sub-surface snow-ice model, surface albedo is calculated from the simulated snow grain shape and size, snow depth, meltwater accumulation, cloudiness and ice albedo. The use of snow metamorphism processes allows a realistic modelling of the temporal variations in the surface albedo during both melting periods and accumulation events. Concerning the surface albedo, the main finding is that an accurate albedo simulation during the melting season strongly depends on a proper initialization of the surface conditions which mainly result from winter accumulation processes. Furthermore, in a sensitivity experiment with a constant 0.8 albedo over the whole ice sheet, the average amount of melt decreased by more than 60%, which highlights the importance of a correctly simulated surface albedo. The use of this coupled atmosphere–snow RCM offers new perspectives in the study of the Greenland surface mass balance due to the represented feedback between the surface climate and the surface albedo, which is the most sensitive parameter in energy-balance-based ablation calculations.     

A comparison of photolysis rate parameters estimated from measured and simulated actinic flux for wintertime conditions at Storm Peak Laboratory,Colorado

Photolysis rate parameters depend upon solar actinic flux and chemical species dependent quantum yields and cross sections. Spectrally resolved measurements of actinic flux should be preferred over flux derived from models for the analysis of field observations. Actinic flux can be difficult to derive from the irradiance measurements of flat-plate radiometers. It is also difficult to estimate from models due to uncertainties in the ozone column, aerosol concentrations and distributions, cloud cover, optical depth and surface albedo. A series of actinic flux measurements were performed at Storm Peak Laboratory (3,210 m above sea level), Colorado, United States with spectroradiometers during the wintertime (January 07–10, 2004). The site is relatively remote with a clean atmosphere and during the wintertime the ground is generally covered by fresh snow with a high albedo. The actinic flux measurements were used to estimate the photolysis rate parameters of ozone, nitrogen dioxide and formaldehyde. The measured actinic flux and the photolysis rate parameters derived from the flux were compared to calculations using the Tropospheric Ultraviolet-Visible Model (TUV), version 4.2 (Madronich and Flocke, 1998). The TUV modeled actinic flux, the measured flux and the photolysis rate parameters derived from them had similar temporal patterns. However there were significant differences in their magnitude due to uncertainties in the data available to initialize the TUV model and the calibration of the spectroradiometer.     

An Intercomparison Among Four Models Of Blowing Snow

Four one-dimensional, time-dependent blowing snow models areintercompared. These include three spectral models, PIEKTUK-T,WINDBLAST, SNOWSTORM, and the bulk version of PIEKTUK-T,PIEKTUK-B. Although the four models are based on common physicalconcepts, they have been developed by different research groups. Thestructure of the models, numerical methods, meteorological field treatmentand the parameterization schemes may be different. Under an agreed standardcondition, the four models generally give similar results for the thermodynamic effects of blowing snow sublimation on the atmospheric boundary layer, including an increase of relative humidity and a decrease of the ambient temperature due to blowing snow sublimation. Relative humidity predicted by SNOWSTORM is lower than the predictions of the other models, which leads to a larger sublimation rate in SNOWSTORM. All four models demonstrate that sublimation rates in a column of blowing snow have a single maximum in time, illustrating self-limitation of the sublimation process of blowing snow. However, estimation of the eddy diffusioncoefficient for momentum (K_m), and thereby the diffusion coefficients for moisture (K_w) and for heat (K_h), has a significant influence on the process. Sensitivitytests with PIEKTUK-T show that the sublimation rate can be approximately constant with time after an initial phase, if K_m is a linear function with height. In order to match the model results with blowing snow observations, some parameters in the standard run, such as settling velocity of blowing snow particles in these models, may need to be changed to more practical values.     

Rapid vegetation responses and feedbacks amplify climate model response to snow cover changes

We investigate the response of a climate system model to two different methods for estimating snow cover fraction. In the control case, snow cover fraction changes gradually with snow depth; in the alternative scenarios (one with prescribed vegetation and one with dynamic vegetation), snow cover fraction initially increases with snow depth almost twice as fast as the control method. In cases where the vegetation was fixed (prescribed), the choice of snow cover parameterization resulted in a limited model response. Increased albedo associated with the high snow caused some moderate localized cooling (3–5°C), mostly at very high latitudes (>70°N) and during the spring season. During the other seasons, however, the cooling was not very extensive. With dynamic vegetation the change is much more dramatic. The initial increases in snow cover fraction with the new parameterization lead to a large-scale southward retreat of boreal vegetation, widespread cooling, and persistent snow cover over much of the boreal region during the boreal summer. Large cold anomalies of up to 15°C cover much of northern Eurasia and North America and the cooling is geographically extensive in the northern hemisphere extratropics, especially during the spring and summer seasons. This study demonstrates the potential for dynamic vegetation within climate models to be quite sensitive to modest forcing. This highlights the importance of dynamic vegetation, both as an amplifier of feedbacks in the climate system and as an essential consideration when implementing adjustments to existing model parameters and algorithms.     

Snow cover validation and sensitivity to CO2 in the UVic ESCM

Abstract

The University of Victoria's (UVic) Earth System Climate model is used to conduct equilibrium atmospheric CO₂ sensitivity experiments over the range 200–1600 ppm in order to explore changes in northern hemisphere snow cover and feedbacks on terrestrial surface air temperature (SAT). Simulations of warmer climates predict a retreat of snow cover over northern continents, in a northeasterly direction. The decline in northern hemisphere global snow mass is estimated to reach 33% at 600 ppm and 54% at 1200 ppm. In the most northerly regions, annual mean snow depth increases for simulations with CO₂ levels higher than present day. The shift in the latitude of maximum snowfall is estimated to be inversely proportional to the CO₂ concentration. The northern hemisphere net shortwave radiation changes are found to be greater over land than over the ocean, suggesting a stronger albedo feedback from changes in terrestrial snow cover than from changes in sea ice. Results also reveal high sensitivity of the snow mass balance under low CO₂ conditions. The amplification feedback (defined as the zonal SAT anomaly caused by doubling CO₂ divided by the equatorial anomaly) is greatest for scenarios with less than 300 ppm, reaching 1.9 at the pole for 250 ppm. The stronger feedback is attributed to the significant albedo changes over land areas. The simulation with 200 ppm triggers continuous accumulation of snow ('glaciation') in regions which, according to paleo‐reconstructions, were covered by ice during the last glacial cycle (the Canadian Arctic, Scandinavia, and the Taymir Peninsula).     

Assessment of GCM simulated snow albedo using direct observations

 Annual cycles of monthly albedos simulated with a general circulation model (GCM) are compared with surface observations. The data observed at 35 stations are retrieved from the Global Energy Balance Archive (GEBA) and drawn from the soil moisture and meteorological observations in the former Soviet Union. The model data are obtained with the ECHAM4 GCM in a ten-year simulation of the present-day climate at T106 resolution. The model calculated albedo values are modified before they are compared with the surface observations: They are interpolated to the stations and adjusted to account for altitude differences and fractional forest area. During the snow-free period, the model underestimates the albedo by up to 0.05 at the stations (with values between 0.2 and 0.25 measured over short grass) because the albedo for grassland is too low in the model. During the period with seasonal snow cover, the model underestimates the albedo by up to 0.2 at stations in Russia, Scandinavia and Canada, which experience severe winters. This underestimation is due to an oversimplified parameterization of the snow covered grid fraction and an inadequate linear relation between snow albedo and temperature. The derivative of albedo with respect to the forest fraction implemented in ECHAM is in line with the observations, although a small overestimation of the model’s gradient has been detected. Received: 3 July 1998 / Accepted: 24 December 1998