BACKGROUND

The surface heat budget of the Arctic Ocean determines the state of the surface. If the budget is positive, the surface is gaining energy. This excess energy can manifest as warming of the water, ice, or snow. When ice or snow finally reach 0 deg C, however, the excess energy goes into melting. In the opposite case, if the surface energy budget is negative, water freezes, then the ice or snow begins cooling.

In turn, the albedo of the surface--and thus the surface's ability to absorb and dispense shortwave radiation--depends on the surface state. Cold, dry snow has an albedo to 0.8-0.9; melting snow or bare sea ice can have an albedo as low as 0.3. Melt ponds have even lower albedos; and open water is almost black, its albedo being less than 0.1. Thus, there is significant interplay between the state of the surface and its radiative properties.

Stratiform clouds are almost ubiquitous in the Arctic summer, undoubtedly because of the presence of surface water in leads, melt ponds, and melting snow. Although these clouds limit the amount of sunlight reaching the surface, the water on the surface makes its albedo lower. The surface thus absorbs a higher fraction of the shortwave radiation reaching it than if it were dry or frozen. These low, relatively warm summer clouds also have a higher effective longwave emissivity than winter clouds and therefore tend to push the surface longwave radiation balance positive.

These feedbacks among the state of the surface, its albedo, and the clouds are a key puzzle in the Arctic. In the above summer scenario, for example, it is not obvious whether the clouds decrease the surface radiation budget by limiting shortwave radiation or increase it through their longwave properties and their close coupling with the surface state.

Directly measuring all the terms in the surface heat budget for all seasons in the Arctic is our method for investigating this close coupling among surface properties, surface albedo, and the characteristics of the atmospheric boundary layer, including its clouds. Following the time series of the components in the surface heat budget tells--among many other things--when the surface should begin melting; when it should begin freezing; when the atmosphere is stably stratified and, thus, when the vertical exchange that fosters clouds should be limited; and when longwave or shortwave radiation dominate the changes in surface state.

APPROACH

Our theme is that it is impossible to understand all the feedback processes coupled through the surface state without detailed measurements of the surface heat budget. As such, our three-year proposal requests funding for continuous surface-level measurements of the various fluxes in the surface heat budget and attendant meteorological variables at the main ice camp and at three remote sites for the entire duration of the SHEBA experiment.

The focal point of our research program is a 20-m tower at the main SHEBA camp. This will have multiple levels of instruments for measuring mean wind speed, direction, temperature, and humidity and three or four levels with fast-responding instruments for measuring turbulent fluxes directly by eddy correlation. Near this tower will also be a variety of instruments for measuring radiative fluxes and surface temperature.

We will deploy similar instruments, but at one level only, at three sites 3-5 km from the main camp. Most of these remote sites will be collocated with the sites of Perovich's ice physics group. In addition, we will have two portable towers with similar instrumentation; these will provide for intercalibration among the various sites and for studying ephemeral features such as leads and melt ponds. One surface-layer scintillometer systems, which propagates a laser beam over a 400-m, horizontal path, will also be deployed near the main camp. The scintillometer measures path-averaged momentum and sensible heat fluxes. The remote, portable, and path-averaging instruments, when combined with the measurements at the main camp, are designed to provide information on areally averaged fluxes. To complement our measurements and to facilitate their scaling up, we will count on the information provided by surface-based remote sensors, remote buoys, aircraft, and satellites funded under other proposals.

SCIENCE QUESTIONS

There are a host of good science questions that we can treat with the data we propose to collect. Here we highlight a few key questions.

Scaling Up, Scaling Down

Arctic sea ice is heterogeneous. That is why we will have a main site and three remote sites. Each will be placed over a different ice type. By knowing the areal fraction that each ice type covers and by measuring the surface heat budget over each ice type, we will investigate methods of scaling measurements at a single point up to the size of a mesoscale grid box. The multiple levels for flux measurements on our main tower will also facilitate this scaling up. Each height will be sampling a different upwind footprint. Flux measurements that are not constant in the vertical reflect the effects of the heterogeneous surface and allow us to test methods for aggregating the flux measurements from the remote sites. The scintillometer is inherently an areally averaging instrument and, thus, will provide an intermediate-scale link between the point measurements and the true wide-area averages.

The opposite side of the scaling up question is scaling down: Is it possible to estimate from the predictions in a GCM or mesoscale model's grid box what the local fluxes might be. When we solve the scaling up problem, the scaling down problem should yield to inversion.

Tracking the Surface Heat Budget

A unique feature of the Arctic ice cover is that ice concentration decreases slowly in the spring but increases relatively fast in the fall. As we mentioned earlier, the time series of terms in the surface heat budget should correlate with observations such as this and with many other observables. Once we solve the scaling up problem, we will construct time series of the areally averaged components in the surface heat budget and investigate correlations between these terms and events such as the onset of the melt season; the onset of freeze-up; and changes in the fractional cloud cover, the ice albedo, the ice thickness, and the temperature of the mixed layer.

Stable Stratification

In the winter, especially, the Arctic atmospheric boundary layer is stably stratified. Stable stratification is a modeler's bane. The turbulent exchanges are intermittent and therefore not handled well by the usual bulk parameterizations. With only intermittent turbulent heat transfer, it is often computationally impossible to keep the predicted surface temperature from falling to unrealistic levels in response to a negative longwave radiation balance. In reality, however, the surface temperature does not fall without bounds. With our surface flux measurements and in collaboration with the Uttal remote-sensing group, we will investigate processes in the stable boundary layer that prevent this runaway drop in surface temperature that models often predict.

We speculate that breaking internal gravity waves in the atmosphere frequently sweep down to the surface bringing warmer air from aloft that keeps the surface somewhat equilibrated with the higher atmosphere. With the radars, lidars, and sodars, we will look for these breaking waves and measure there contributions to the surface heat budget with our surface sensors.

Parameterizing the Turbulent Surface Fluxes

Andreas developed a theoretical model for parameterizing the turbulent surface fluxes of sensible and latent heat over snow-covered sea ice. Parameterizations of this type are common in one-dimensional and mesoscale models because they allow estimating the turbulent surface fluxes from routine measurements. Although Andreas's parameterization has been partially verified over snow-covered ground and glaciers, it has not been validated over sea ice, though it is the only one that has been specifically developed for sea ice. Our measurements will provide data for checking this parameterization and revising it if necessary.

What Happens in the Summer?

Most surface-level measurements over sea ice that have been reported in the open literature were made in the winter, when the surface snow is dry and erodible. The summer is largely unknown and definitely under-reported. In the summer, the snow is wet or gone and, thus, does not drift as it does in winter. The surface is covered with melt ponds that keep the surface-level air near saturation (that's why all the clouds) and give the surface a low albedo. The interplay between the morphology and the wetness of the surface and its resulting dynamic characteristics make the summer quite interesting. This interplay is also the crux of the ice-albedo feedback problem.

We expect the summer ice surface to respond to the wind somewhere between the way the open ocean and an asphalt parking lot would. That is, with few ponds, the surface is immobile and would act like a flat plate. With a lot of ponds, the higher the wind, the higher the waves in the ponds, and the rougher the surface. Thus, as over the ocean or a field of waving wheat, the dynamic properties of the surface are directly related to the wind speed. Likewise, because the roughened water in melt ponds has a different albedo than still water, in summer the turbulent characteristics of the atmospheric boundary layer are closely tied to surface albedo. Our data, the data that the ice physics group will collect, and satellite observations will make a very good set for sorting out how surface albedo in the summer responds to the meteorology.

COLLABORATION/COORDINATION

We will rely on data to complete our set from several other SHEBA groups. In turn, we have also agreed to collaborate experimentally or to provide our data to other SHEBA programs. Here is a partial list of that coordination.

Perovich's Ice Physics Group

We will collocate our three remote sites with this group's sites. We will provide them routine meteorological data; our two groups will collaborate on measuring the radiative fluxes; and they will provide us in-snow and in-ice temperature profiles that we need for computing the subsurface conductive flux that closes the surface heat budget. Collaboratively, we will investigate how changes in surface morphology (and consequently albedo) correlate with the components of the surface heat budget.

Paulson's Summer Lead Study

Using our two portable flux towers, we will collaborate with the summer lead study of Paulson's group. We will provide turbulent and radiative flux measurements and high quality surface-level meteorological quantities upwind and downwind of leads. These measurements will provide data and parameterizations for quantifying the energy available at the surface of the lead. Paulson's in-water measurements will evaluate how this energy gets distributed in the mixed layer.

Uttal's Surface-Based Remote Sensing Group

Fairall is also a collaborator on this proposal, so there will be close coordination. In particular, we will rely heavily on the remote sensors to help us characterize the turbulent and cloud properties of the atmospheric boundary layer and the local forcing above the boundary layer. We will coordinate our surface sampling and analysis with remotely sensed evidence of internal waves to investigate our hypothesis that these waves are a mechanism that keeps the surface from cooling catastrophically in the presence of a large, negative longwave radiation balance.

Curry's and Brooks' Aircraft Observations

Both aircraft measure turbulent fluxes and some of the radiation components. Curry plans to map the surface in the SHEBA region, including ice type, pond coverage, and open water fraction. Data from both these platforms will thus be useful when we try to scale up our surface-based flux measurements. In turn, the aircraft measurements, especially Curry's, will be too high to yield the actual turbulent surface fluxes. Our flux measurements at the several surface sites will provide ground truth and facilitate extrapolating the aircraft flux measurements down to the surface, thereby generating flux profiles. Again, this capability of areally averaging surface fluxes is crucial for understanding how the surface state responds to the forcing.

Overland's Buoy Array

This large buoy array will provide surface-level pressure fields and routine meteorological data that will place the SHEBA camp in a larger scale synoptic context. These pressure fields will also provide estimates of the geostrophic wind. By comparing these geostrophic winds with our measurements of the surface momentum flux, together we will be able to evaluate geostrophic drag coefficients. GCMs commonly use such coefficients to infer surface wind and surface stress from computed pressure fields. By developing turbulent flux parameterizations on the basis of our main-camp observations, we also will be able to use these far remote data to aggregate local observations up to even larger scales.

In summary, our program, Uttal's program, and this one approach the aggregation problem as if it were a matrioshka, a nested Russian doll. Our measurements, the smallest doll, will yield the direct results of the atmospheric forcing--the mean local conditions and the fluxes of heat and momentum. Uttal's program, which will provide the local forcing that drives these fluxes, is the doll that surrounds ours. And Overland's array is the next larger doll--the geostrophic forcing.

Lindsay's and Krueger's Modeling Proposals

We will provide our basic data to Lindsay for his modeling work. With our participation in the summer lead study, we will support Krueger's lead modeling.