In the global climate system, clouds are arguably the biggest regulator of radiation. Thick, bright clouds reflect energy from the Sun during the day, causing Earth’s surface to be cooler than on a cloudless day. High, thin clouds allow energy from the Sun to reach the surface while trapping heat emitted by the planet, thus acting as a blanket and causing the surface to be warmer than on a cloudless day. And at night, with no incoming solar radiation to reflect, clouds trap heat and warm the surface.
We can infer from those simple examples that the physical characteristics of clouds, as well as the times and places they appear, profoundly affect surface conditions.
Although clouds and their impact on solar radiation have been recognized, measured, and modeled over much of the globe for decades, their behavior in the polar regions has, until recently, remained largely a mystery. Detailed, ground-based cloud observations are biased toward populated areas that have consistent access to electricity to power meteorological instruments. Isolated flight and ship campaigns have taken place in the Arctic, and a handful of permanent scientific stations exist in both polar regions. However, measurements of polar cloud properties from the ground have been regionally sparse, temporally sparse, or both.
Much of the satellite cloud observation record consists of measurements made by passive instruments, which simply measure radiation already coming off the planet, much like a camera without a flash. The instruments require either reflected incoming solar illumination or strong contrasts in outgoing IR radiation to operate. In polar regions, the winter darkness—along with conditions such as cold clouds overlying snow, blowing snow, and ice—makes extracting meaningful data on cloud behavior from the satellite record difficult. Continuing the analogy above, the task is like trying to discern details in the picture taken in the dark without the flash.
Yet an accurate picture of the surface energy balance and the role clouds play in modulating it is especially and increasingly necessary in polar regions because of ice melt. Understanding the mass balance of the Greenland and Antarctic ice sheets, as well as Arctic sea ice loss, are high priorities in the climate community for many reasons—among them, habitat loss, sea-level rise, and impacts on the overall ocean and atmosphere circulation. Modeling and predicting changes in ice melt accurately requires tools that can reliably measure polar cloud distribution and behavior.
CloudSat and CALIPSO were launched by NASA in 2006, specifically to improve understanding of clouds and climate. The satellites respectively carry a 94-GHz Cloud Profiling Radar (CPR) and a two-channel lidar, which provide detailed vertical structure of clouds and precipitation. Much like the weather radars we are all familiar with, CPR actively sends out signals and measures their reflectivity. Likewise, lidar sends out a pulse of visible light and measures the amount of light scattered back to the instrument. Neither sensor requires solar illumination to operate, making them ideal for increasing our understanding of polar regions. Continuing with the camera metaphor, the two satellites now provide our flash, and the pictures they take in the dark are full of details we were previously unable to see.
Figure 1. NASA and its international partners maintain the Afternoon Constellation, or A-Train for short. The constellation's satellites follow each other as they orbit Earth. CREDIT: NASA
CloudSat and CALIPSO joined the A-Train constellation of satellites, where their constant co-location with other instruments allows for detailed studies of cloud phase, cloud water content, precipitation type and rates, and more. (To learn more about the constellation, see Tristan L’Ecuyer and Jonathan Jiang's article, "Touring the atmosphere aboard the A-Train," which appeared in Physics Today's July 2010 issue.) Thanks to the combined measurements, its possible for the first time to calculate radiative fluxes directly from observations throughout the full atmospheric column. The measurements span nearly the full globe (82°N to 82°S), completing one orbit cycle approximately every 90 minutes and completing a full Earth observation cycle every 16 days.
These flux calculations require many inputs, including detailed cloud properties (from CloudSat’s CPR, CALIPSO’s CALIOP, and Aqua’s MODIS instruments), aerosol properties (from CALIOP), surface characteristics (from MODIS and Aqua’s AMSR-E) and temperature and humidity profiles (from ECMWF reanalysis). All of those inputs are used to initialize a broadband radiative transfer model, which in turn provides the computed flux values. The values tell us the quantitative impact that clouds have on the radiative balance at the surface.
Traditionally, there are two radiative impacts considered for clouds: cloud longwave forcing (CLWF) and cloud shortwave forcing (CSWF). Radiative forcing is broken into the two categories because of their competing effects on the radiative balance. CSWF, as the name implies, measures the impact that a cloud has on shortwave energy (λ < 4 μm). There are nuances beyond the scope of this discussion, but the general impact that a cloud has on shortwave radiation is to reflect incoming solar radiation back to space, causing the surface to receive less radiation and to cool. (You can feel this phenomenon when a cloud passes overhead on a warm day.)
CLWF measures the impact a cloud has on the longwave energy (λ > 4 μm) emitted and absorbed by the Earth’s surface and throughout the atmospheric column. Again there are many nuances, but the general effect of a cloud’s presence in the longwave is to trap heat and warm the surface, reducing the radiation emitted to space. This is why cloudless nights are generally cooler than cloudy nights in a given location and season.
The magnitudes and details of the two competing cloud effects depend on the surface type as well as many cloud properties, including phase, particle size, temperature, height, water content, and thickness. Alone, each forcing provides important insight into the impact of a cloud on the energy balance of a column, but their ratio (CSWF/CLWF) clearly shows the overall warming or cooling effect of a particular cloud upon the surface.
Figure 2. Observed Cloud Impact on Surface Radiation Ratio (CISRR) in the Arctic. Data are shown for the months of March, May, July, and September. Each panel represents an average of all satellite overpasses for that month from the years 2007–10. Blue regions show where the cloud shortwave forcing dominates and the average cloud cover serves to cool the surface. Red regions show where the cloud longwave forcing dominates and the average cloud cover serves to warm the surface.
Monthly averages of this ratio, called the Cloud Impact on Surface Radiation Ratio (CISRR), indicate that there are large spatial and temporal variations of cloud impact over the summer months in the polar regions. In figure 2 regions in red indicate where CISRR has a value less than 1.0, meaning the CLWF is dominant and the clouds warm the surface. Regions in blue indicate where CISRR has a value greater than 1.0, meaning the CSWF is dominant and the clouds cool the surface.
CISRR characterizes variations of cloud impact over both the Arctic and the Antarctic regions. The following examples from the Arctic demonstrate the kind of insights that can be gained from this ratio. CISRR in Antarctica has also been mapped and analyzed, but I will omit that data here in the interests of brevity.
In March (upper left of figure 2) the Arctic winter season has just ended. Although there is incoming solar radiation, the surface is still mostly frozen and therefore highly reflective. Clouds do not reflect significantly more radiation than the reflective, frozen surface does, so in this month clouds primarily serve to warm the surface. Only over the north Atlantic, which is comprised of mostly open water year-round, do clouds have a different impact on the surface. Open ocean reflects very little incoming solar radiation, so the presence of a reflective cloud increases the CSWF and results in a cooling of the surface.
In May (upper right of figure 2) it is clear that overall clouds warm the ice sheets, but to a lesser magnitude than earlier months. They also begin to slightly cool the Eurasian continent and Alaska. By July (lower left of figure 2), clouds strongly cool the entire Arctic region, save for Greenland and the remaining Arctic sea ice. September (lower right of figure 2) shows that the clouds have an approximately neutral effect (neither heating nor cooling the surface) over much of the Arctic, but with strong warming over Greenland and the remaining ice sheet. During Arctic winter months, as previously mentioned, there is no incoming solar radiation, and thus no CSWF. Clouds serve only to warm the surface in these months.
CISRR calculations in the polar regions are possible only because of measurements made by CloudSat and the rest of the A-Train instruments. Co-located global measurements of cloud properties and the insights they provide are key to improving our understanding of current melting rates in the polar regions. Increased understanding of polar melting will improve current global climate models (GCMs) and their ability meaningfully to predict future scenarios. My colleagues and I will soon compare CISRR to GCM results to determine if the current polar cloud physics in models produces the correct radiative balance.
Figure 3. Team of scientists repairing Minna Bluff Automatic Weather Station in Antarctica. From left: University of Wisconsin graduate student Elin McIlhattan, Antarctic Meteorological Research Center scientists Lee Welhouse, Dave Mikolajczyk, and Carol Costanza.
As new and important as this data set is, it is important to keep in mind that it is only one piece of the polar climate puzzle. In the satellite community, satellite data are often taken as ‘truth’, while in the modeling community, model results are often trusted over measurements. The physical reality is likely somewhere in between. It is crucial, therefore, that ground-based observation networks continue to be improved and that flight campaigns continue to be conducted in order to validate the global datasets from both models and satellites.
One example of such efforts is the Antarctic Meteorological Research Center (AMRC) and their Automatic Weather Station (AWS) program. The AMRC installed and maintains a network of nearly 60 AWS over the Antarctic continent (Figure 3). The record of quality-controlled surface measurements, which extends for more than 30 years, provides an invaluable resource for the satellite and modeling communities to validate their data. Our future research with CISRR calculations will include a comparison of satellite flux data to the AWS network’s ground-based measurements.
Elin McIlhattan is a graduate student at the University of Wisconsin–Madison, where she studies polar cloud properties using satellites and models under the direction of Tristan L’Ecuyer. She recently joined the Antarctic Meteorological Research Center (AMRC) for a field season in Antarctica, where she repaired and installed Automatic Weather Stations under the direction of Matthew Lazzara.
