Operation IceBridge data brings new twist to sea ice forecasting

Dec 19, 2012
This is a Digital Mapping System (DMS) mosaic of Arctic sea ice. The dark areas are leads, or open areas of water. Identifying leads is one of the necessary steps in preparing IceBridge’s quick look sea ice thickness data product. Credit: Credit: NASA / DMS team

Shrinking Arctic sea ice grabbed the world's attention again earlier this year with a new record low minimum. Growing economic activity in the Arctic, such as fishing, mineral exploration and shipping, is emphasizing the need for accurate predictions of how much of the Arctic will be covered by sea ice. Every June, an international research group known as the Study of Environmental Arctic Change (SEARCH) publishes a summary of the expected September Arctic sea ice minimum known as the Sea Ice Outlook. The initial reports and monthly updates aim to give the scientific community and public the best available information on sea ice.

Researchers rely on models that use estimated ice thickness data and simulated atmospheric conditions to forecast how will change during the summer. For the first time, near real-time ice thickness data obtained by NASA's Operation IceBridge has been used to correct a forecast model's initial measurements, which could lead to improved seasonal predictions.

In a paper published last month in the journal , Ron Lindsay, IceBridge science team member and Arctic climatologist with the Polar Science Center at the University of Washington in Seattle, outlined efforts to use IceBridge data to improve the accuracy of seasonal sea ice forecasts. Lindsay and colleagues used a new quick look sea ice data product that IceBridge scientists released before the end of the Arctic campaign earlier this year. The quick look data, intended for use in time-sensitive applications like , supplements the final sea ice data product typically released roughly six months after the campaign. By using new data processing techniques, IceBridge scientists were able to publish the quick look measurements in a matter of weeks. "The idea was to make the data available for anyone to use for the Sea Ice Outlook," said sea ice scientist Nathan Kurtz of NASA's Goddard Space Flight Center in Greenbelt, Md.

The work outlined in Lindsay's paper marks the first use of IceBridge quick look data in an ensemble sea ice forecast (computer) model. "An ensemble forecast is where you run a single forecast model many different times," said Lindsay. In this case, they ran the Pan-Arctic Ice Ocean Modeling and Assimilation System (PIOMAS) model seven times using conditions from previous summers. PIOMAS uses sea ice extent, the area of sea containing sea ice, and atmospheric data to simulate ice and ocean conditions.

IceBridge data and thickness measurements made by the Seasonal Ice Zone Observing Network (SIZONet), a multidisciplinary project aimed at observing , served as a way to correct initial sea ice conditions. These initial measurements come from running the with historical atmospheric conditions. Lindsay and colleagues used IceBridge and SIZONet data to adjust these measurements and then used the PIOMAS model to create a forecast of September's mean sea ice extent. To make sure what effect the corrected measurements had, they also ran the model as normal, something known as a control run.

Sea ice near Thule, Greenland. Only about 10 percent of sea ice is above the ocean’s surface, which makes measuring ice thickness challenging. More accurate thickness measurements could potentially improve future seasonal sea ice forecasts. Credit: Credit: NASA / Nathan Kurtz

Careful Measurements

Before this forecasting work could begin though, the researchers had to gather and process data, something that takes the hard work of many people. During March and April of 2012, IceBridge gathered sea ice thickness data using four different airborne science instruments.

First, researchers measure the surface freeboard, or the amount of ice and snow above the sea level height, using a laser altimeter known as the Airborne Topographic Mapper (ATM). Next, they use snow thickness data derived from airborne snow radar and subtract that to get an accurate ice freeboard measurement. This figure is then combined with known average density measurements to calculate total ice thickness, of which freeboard is typically only 10 percent. One other instrument, the KT-19 temperature sensor, was used to detect leads, or openings, in sea ice, which are used to determine the sea level height.

SIZONet scientists used a different method, measurements from a helicopter-borne electromagnetic sensor that detects differences in how well sea ice and ocean water conduct electricity, giving a distance between the sensor and ocean water below.

Collecting measurements is only the beginning of the work. Measurements from the ATM laser have to be combined with information from the aircraft's GPS and inertial navigation systems and the readings have to be filtered to remove things like false returns from low clouds and fog. Preparing instrument data for release is a labor-intensive and time-consuming process that normally takes six months. With the quick look product, it was done in a matter of weeks.

Producing quality data so quickly is challenging, but the process proved a good test of the instrument team's talents. "We gained some valuable insights into our capabilities," said ATM senior scientist John Sonntag. "This new confidence in the quick data may open new avenues for us in the future."

Looking Forward

The September mean ice extent for the corrected model were slightly closer to the actual result than the control forecast run, but both were fairly far off from the actual record minimum. This may have been due to unusual weather over the summer, including a large Arctic storm in August, or to deficiencies in the model simulation of the new very thin ice conditions of the Arctic. Lindsay said winds have a bigger impact on the thinner ice of recent years than on thick ice. It may be possible to redo this experiment, using this summer's atmospheric conditions in the forecasts. "This would tell us the impact of the observations for the weather we actually experienced," said Lindsay.

As a step in a new direction, the study and quick look data collection could improve sea ice forecasts in the future. Providing near real-time sea ice data may also help in other areas, such as evaluating model performance.

With plans to produce another quick look product in the coming 2013 Arctic campaign, Kurtz is hopeful that IceBridge data will be useful to sea ice forecasters and other researchers. "The question is how will people use it," Kurtz said.

Explore further: Famine in the Horn of Africa (1984) was caused by El Niño and currents in the Indian Ocean

More information: www.arcus.org/search/seaiceoutlook/index.php

add to favorites email to friend print save as pdf

Related Stories

Arctic ice gets a check up

Mar 31, 2011

Scientists tracking the annual maximum extent of Arctic sea ice said that 2011 was among the lowest ice extents measured since satellites began collecting the data in 1979. Using satellites to track Arctic ice and comparing ...

NASA's IceBridge seeking new view of changing sea ice

Aug 28, 2012

This year scientists working on NASA's Operation IceBridge, a multi-year airborne science mission to study changing ice conditions at both poles, debuted a new data product with the potential to improve Arctic ...

New method to estimate sea ice thickness

Mar 05, 2008

Scientists recently developed a new modeling approach to estimate sea ice thickness. This is the only model based entirely on historical observations.

Declining sea ice to lead to cloudier Arctic: study

Mar 31, 2012

Arctic sea ice has been declining over the past several decades as global climate has warmed. In fact, sea ice has declined more quickly than many models predicted, indicating that climate models may not be correctly representing ...

Looking deeper into sea ice

Oct 08, 2012

(Phys.org)—This year, satellites saw the extent of Arctic sea ice hit a record low since measurements began in the 1970s. ESA's SMOS and CryoSat satellites are now taking a deeper look by measuring the ...

Recommended for you

NASA sees zombie Tropical Depression Genevieve reborn

10 hours ago

Infrared imagery from NASA's Aqua satellite helped confirm that the remnant low pressure area of former Tropical Storm Genevieve has become a Zombie storm, and has been reborn as a tropical depression on ...

Wave energy impact on harbour operations investigated

14 hours ago

Infragravity period oscillations—waves that occur between 25 and 300 seconds with a wavelength between 100m and 10km—can have an impact on berthing operations, depending on a harbour's geometry.

User comments : 1

Adjust slider to filter visible comments by rank

Display comments: newest first

eachus
5 / 5 (1) Dec 19, 2012
As a statistician, I got sick to my stomach reading this report. I don't know where to begin. Well, let's start with what they are calling a control run. This was a run of PIOMAS with actual Sea Ice Outlook data as the input. I suspect it is a bad model, rather than GIGO (garbage in, garbage out). Since running PIOMAS with the data as corrected by IceBridge, resulted in an almost equally bad prediction, the a statistician would tell you that IceBridge is not worth the money. The right way to proceed is to create a new model, test it against historical data, then validate the new model by making new predictions and seeing if they are accurate. If not? The (virtual) bit bucket with PIOMAS in it has room for an infinite number of failed models.

The PIOMAS model may have cost millions of dollars and taken years to develop. Sucks for the people who may have expected publications and/or degrees from their work. But that is science. Trying to "fix" failed models is quackery.