The role of surface albedo feedback in climate variability and climate change (Invited Presentation)

The presence of snow and ice at earth's high latitudes is generally thought to provide a positive feedback to earth's climate: if a warm anomaly occurs, snow and ice retreat, decreasing surface albedo, increasing net incoming solar radiation, and increasing the magnitude of the original temperature anomaly. (Opposite reasoning applies for cold anomalies.) In this study, this idea is tested by comparing a coupled ocean-atmosphere model simulation where surface albedo feedback (SAF) is artificially suppressed by prescribing surface albedo to one where snow and ice anomalies are allowed to affect surface albedo, as the model was originally designed.

One goal is to see whether SAF truly is positive in the context of the model's internally-generated temperature variability. Surprisingly, in the northern hemisphere (NH), SAF hardly amplifies internal variability, except during springtime, when there is about 20% more variability in mid to high latitudes when surface albedo feedback is present. This is because NH albedo anomalies are mainly generated by anomalies in snow cover. During winter and fall, snow cover is controlled by accumulation, a process associated with the warm temperatures of the precipitating sector of synoptic-scale waves. This results in a positive correlation between snow accumulation and local temperature, effectively disrupting the positive SAF loop. In spring, melting controls the snow budget. Since melting is highly correlated with warm temperatures, this sets up perfect conditions for a large positive feedback. Like snowmelt, the growth and decay of sea ice are directly associated with coldness and warmth, respectively, so that SAF is also strongly positive in regions where sea ice generates most surface albedo variability. For example, in the mid to high latitudes of the southern hemisphere (SH), there is 50-80% more variability in the simulation with SAF during all seasons.

Another goal is to see how much SAF controls the geographical distribution of the warming when greenhouse gases increase. To examine this issue, canonical CO₂-doubling experiments where performed with both models and the resulting equilibrium climates were compared. Though there is about twice as much warming at mid to high latitudes of both hemispheres when SAF is present, there is also a surprisingly large degree of polar amplification in the experiment without SAF, with the poles warming approximately twice as much as the tropics. This occurs because of a significant increase in sensible heat flux from the warm ocean to the cold atmosphere when sea ice thickness and extent are reduced in the warmer climate. Another surprise from these experiments is the effect of SAF on the warming in the tropics. In this region, there is approximately 20% more warming when SAF is present, in spite of the absence of any local change in simulated surface albedo. This occurs because of the larger decrease in the equator-to-pole temperature gradient in the experiment with SAF. Since the poleward heat transport is sensitive this gradient, a larger reduction in the poleward heat transport occurs when CO₂ is doubled in the presence of SAF, resulting in more tropical warming when SAF is present.