"If you just look at one or two days the differences [in accuracy between the superensemble and conventional forecasting methods] may not be that large, but if you extend this to four or five, six days then you can see a huge difference," Krishnamurti said. "The superensemble seems to outperform most models for longer ranges. Under shorter range, the differences are smaller, but still it's probably one of the best models that exists today."
Krishnamurti and his team of researchers at FSU's Real Time Hurricane Forecast Center use up to 11 tropical forecasts from weather and climate models around the world. The team uses the forecasts with observational data to derive some statistics on the behavior of the models, then removes biases from the ensemble as a whole. The result is a more accurate one- to six-day track and intensity forecast.
"We have seen important and somewhat similar applications in the areas of global weather and seasonal climate prediction," said Eric Williford (FSU). "We believe that substantial improvement to these areas can be made using the superensemble technique."
The researchers are now working on experimental real-time Atlantic hurricane prediction, including 1999's hurricanes Bret and Dennis. The superensemble method was able to predict both storms' paths very accurately. This was done using FSU's forecast models, models from around the world, and previous Atlantic storm data. The more information researchers provide to the computer about previous storms, the more accurate the prediction seems to be.
"There's a lot of scope for improving the research so we can keep improving what we are doing," Krishnamurti said. "We need to do this on a much finer grain so we can compute these forecasts more accurately and tell where there will be heavy rains and landfall during hurricane events."
David Morse (University of Texas at Austin) said the project has implications for scientists seeking to predict changes in sea level related to the gradual shrinking of the West Antarctic ice sheet. However, the discoveries are noteworthy in their own right because they reveal new information about the rivers of ice that flow across the continent.
Antarctica's two ice sheets store more than 90% of the earth's fresh water. The West Antarctic ice sheet contains enough water to raise sea level between 4.5 and 6 meters (15 and 20 feet). The ice sheet is a dynamic structure flowing continuously toward the sea. Its flow is organized into ice streams--rivers of ice that are tens of kilometers wide and hundreds of kilometers long, flowing as much as 100 times faster than the surrounding ice. They are so unlike the surrounding ice that they are easily seen from space.
Morse worked with a group of scientists reviewing measurements of ice sheet movements collected for the first time in Antarctica by satellite. Ice velocity previously had been measured by pounding a stake into the ice and taking measurements as the stake moves slowly along with the ice itself. Morse said that satellite data allow a continuous measurement of velocities over large enough areas to give scientists a much clearer picture of the physics of ice movement.
The satellite information was examined by scientists at NASA and the Canadian Space Agency to produce a new map of ice surface velocities in a largely unexplored region further inland than previously identified ice streams. Combining this information with ground-based velocity observations made by NASA scientists as well as ice thickness measurements collected by the airborne geophysics group at UT Austin, the scientists made three important discoveries. First, they found proof that the ice streams are fed by an extensive network of previously undiscovered tributaries that extend hundreds of kilometers further upstream than expected. Second, they discovered that the ice streams are linked through a system of shared tributaries, making predictions about their movement much more challenging. Third, scientists found that the tributaries coincide with valleys below the ice. "This is significant since it gives us insight for predicting where ice streams might have occurred in the past or may occur in the future," Morse said.
The study was financed by several agencies, including NASA, NSF, and the Canada Center for Remote Sensing. Results were published in Science. The measurements were part of a project to produce a high-resolution satellite map of the entire continent.
The phenomenon affects climate locally, and probably regionally, according to researchers from the Consilio Nazionale delle Ricerche in Bologna, Italy, and the University of Washington (UW), who reported their findings in Nature.
Such organic particles bring a decrease in surface tension--commonly observed when a small amount of soap or detergent is added to water--during the formation of cloud droplets. The result is the formation of more droplets than normal, making clouds that are somewhat more reflective of sunlight, said Robert Charlson, a UW atmospheric chemist. Charlson coauthored the study with Maria Cristina Facchini, Mihaela Mircea, and Sandro Fuzzi of Consilio Nazionale delle Ricerche.
Different organic substances have different effects on surface tension, depending on the molecular nature of the particles, he said. "Just compare a teaspoonful of sugar to a teaspoonful of detergent in, say, a bucket of water. The sugar does virtually nothing while the soap makes the water sudsy," Charlson said.
To reach their findings, the researchers collected large volumes of water from clouds in the Po River Valley of northern Italy. The water then was evaporated to simulate the reverse process of droplet formation.
"We found out that the stuff in the Po valley is more like soap than sugar," Charlson said. The researchers discovered up to a 30% decrease in surface tension in the cloud droplets, which they calculate would result in about 20% more droplets than normal and reduce the size of droplets about 6%.
At the top of the atmosphere, the increase in the number of droplets results in about 1% more sunlight being reflected away from earth, they said. The effect of the small increase in reflection is a negative feedback to global warming, but the pollution particles also represent a fundamental change in the chemistry of the atmosphere. Similar effects are likely in other agricultural and industrial regions, Charlson said, but no data are available to make a global estimate.
The researchers noted that many organic compounds that turn up as atmospheric particles, or aerosols, often are listed as insoluble but in reality are soluble under the conditions that trigger droplet formation.
The study also indicated flaws in the yardstick by which gases are compared under the Kyoto Protocol, an agreement now signed by 84 countries, including the United States, that was negotiated in December 1997 with the intent of slowing global warming.
"The main finding is that including gases other than CO2 emissions from fossil fuels could greatly reduce costs of meeting the protocol," said John Reilly, lead author of the paper and associate director for research at the MIT Joint Program on the Science and Policy of Global Change. "Economically efficient policies will be required that encourage reduction of these emissions--not an easy task, as reductions must come from sources as diverse as landfills, aluminum production, livestock, and electrical switchgear."
Currently, most policy discussions and economic analyses narrow climate issues to a debate about emissions of CO2 from fossil fuels. According to the new study, this focus on CO2 alone has led to an overestimate of approximately 21% in predicting annual costs in 2010 for meeting Kyoto Protocol emissions caps in industrialized regions.
The new study also considered atmospheric interactions among the other greenhouse gases, climate feedbacks, the roles of carbon monoxide and nitrogen oxides (key components of smog), and aerosols' cooling effect. It showed that if non-CO2 gases are included in the Kyoto Protocol, the same reduction in warming can be achieved much more cheaply than if only CO2 had been included. Achieving the reduction by controlling fossil CO2 only could cost over 60% more than an effort controlling other gases as well.
The Nature study also notes flaws in the yardstick used to compare greenhouse gases under the Kyoto Protocol. The protocol calls for reductions in emissions of several greenhouse gases to be credited against a CO2-equivalent emissions cap. This cap is calculated in terms of global warming potential (GWP), an index defining the contribution of each greenhouse gas to atmospheric warming relative to CO2. The Nature article shows that use of GWPs as applied in the Kyoto Protocol mitigates climate change considerably more for multigas strategies than for supposedly equivalent CO2-only control when emissions cuts are deep enough to stabilize the radiative effects of these gases. Differences in effects between strategies develop early in the 21st century, soon after the point at which the researchers hypothesize that the protocol will be extended to include developing countries, in which emissions of non-CO2 greenhouse gases are substantial.
Additional authors of the Nature paper are Jochen Harnisch, Jean Fitzmaurice, and Henry Jacoby of MIT; David Kicklighter and Jerry Melillo of the Marine Biological Laboratory, Woods Hole; and Peter Stone, Andrei Sokolov, and Chien Wang of MIT. The research was funded by the MIT Joint Program on the Science and Policy of Global Change.