An overview of projects throughout the organization
Drought destroyed the corn crop of this farmer in Lisutu, Zambia, in 2002. New analyses from NCAR and NOAA suggest that drought may intensify across southern Africa. (Photo © 2002 Richard Lord/UMCOR.)
Climate change in Africa
CGD's Jim Hurrell and Adam Phillips, together with NOAA scientists Marty Hoerling and Jon Eischeid, recently drew from 80 simulations of global climate produced by five computer models, including NCAR's Community Atmosphere Model, to show how ocean warming consistent with the build-up of greenhouse gases in the atmosphere will affect Africa's climate.
The computer models simulated observed drought conditions in southern Africa that have occurred in response to a 0.6 degree Celsius (1 degree Fahrenheit) warming in the Indian Ocean since 1950. According to the analysis, continued warming of the Indian Ocean will cause drought to intensify in coming decades over the region, which has been experiencing recurrent drought since the 1970s. As showers and thunderstorms develop in the rising air above the warming ocean, the result is sinking air that causes drought in southern Africa.
"In our models, the Indian Ocean shows very clear and dramatic warming into the future, which means more and more drought for southern Africa," Jim says. "The observed Indian Ocean warming is consistent with what the models indicate as a response to an increase in greenhouse gases, and this effect only becomes stronger in future decades."
The opposite might be the case farther north, according to the analysis. Most models show the Sahel region of Africa, which stretches from the Atlantic coast to the Horn of Africa, experiencing stronger monsoons in the future as the North Atlantic Ocean heats up and produces moister conditions over the region. For decades, the Sahel suffered catastrophic droughts that Jim and colleagues attribute to a largely natural cooling of the North Atlantic Ocean during the late 20th century. For more information about the study.
Wildfires and ozone
Scientists have long known that wildfires are a significant source of tropospheric ozone. But preliminary findings by ACD's Frank Flocke and Aaron Swanson, who is now at Northrop Grumman Corporation, indicate that, under some conditions, the fires may result in less ozone than commonly believed. If confirmed, this would have implications for global and regional models that seek to quantify levels of the pollutant, which is formed by sunlight acting on hydrocarbons and nitrogen oxides.
Frank and Aaron used a new instrument, developed at NCAR and the Georgia Institute of Technology, that measures PAN (peroxyacetyl nitrate). PAN is a useful tracer for tropospheric ozone because both chemicals form in the atmosphere from the same ingredients.
During an experiment in New England last year, Frank and his colleagues at NOAA's Aeronomy Laboratory in Boulder took measurements of chemicals in a plume that had reached the East Coast from forest fires in Canada and Alaska. The plume contained somewhat elevated levels of ozone. But the ratio of PAN to ozone was only about 1-to-20, which was much lower than the expected ratio of 1-to-30 or more.
Frank and his colleagues suspect that chemical reactions involving soot and other particles in wildfire plumes may reduce the production of ozone. ACD's Louisa Emmons, Peter Hess, Jean-Francois Lamarque, and Gabriele Pfister will work with Frank on computer modeling to probe deeper into the processes.
The image at right scores better on the Critical Success Index and other traditional measures of forecast skill, because rain (area in O) was correctly forecast (area in F) at least some of the places that it fell. The forecast at left has a score of 0 (no skill) even though it depicted the size and shape of the rain area more accurately, and would be considered a good forecast from many perspectives. (Image courtesy Barb Brown.)
Testing the skill of forecast models
Staffers in RAL and MMM are collaborating on a new technique, called object-based verification, that should allow scientists to test the skill of today's fine-scale forecast models more thoroughly and meaningfully. Last month, RAL's Barb Brown and MMM's Chris Davis gave seminars at NCAR on the technique. They are working with RAL's Randy Bullock and Daran Rife through support from the Federal Aviation Administration, NSF, U.S. Army Test and Evaluation Command, and the U.S. Weather Research Program.
Traditional methods for examining forecast skill focus on fixed points in space: whether it rained or not at a weather station, for instance. But these techniques tend to reward overforecasting, so a prediction for isolated showers could be essentially correct and still rate poorly in a statistical sense. This long-recognized problem gets worse as the resolution in forecasting models goes up. "People are unhappy that the detail in higher-resolution models is punished by the standard methodologies," says Barb.
Instead of focusing on points on the ground, object-based approaches treat the weather itself as an object: an area of rain, a zone of intense wind shifts, and so on. Each object is identified and tracked by software that recognizes patterns, taking into account the object's size, intensity, orientation, and even how smooth or jagged its borders are. By focusing on objects, the new technique evaluates the accuracy of a forecast by capturing the character of a weather feature.
Barb and Chris have been using data from the Weather Research and Forecasting model (WRF) to try out the new approach. For the summers of 2001 and 2002, Barb and colleagues identified about 48,000 objects in the model, and they found matches in precipitation data for more than 70% of them. "This high number of objects allows for quantitative analysis with high statistical value," Barb says. Chris is adding the variable of time, looking for errors in how early or late WRF caught on to weather features in BAMEX, the 2003 Bow Echo and MCV (Mesoscale Convective Vortex) Experiment.
"Defining objects allows us to get closer to real phenomena," Chris says. According to Barb, "There are a lot of traditionalists in the verification world, but I think they're coming around to this kind of approach."
Workshop on integrated weather studies
A two-part workshop called WAS*IS (Weather and Society * Integrated Studies) will begin in November. Sponsored by the Societal Impacts Program, which is part of ISSE and RAL, the workshop will explore how weather information can be better communicated to help stakeholders. The organizers, Eve Gruntfest and Julie Demuth, both visiting scientists in ISSE, hope to change from what WAS to what IS the future of integrated weather studies.
One of the main objectives of the workshop is to identify a cadre of early-career professionals dedicated to integrating weather research and social impacts. This will ensure that social impact efforts are an integral part of new project development by scientists and practitioners.
The workshop will emphasize communication strategies and evaluation techniques. Key concepts include
• developing strategies to train physical and social scientists to better understand the fundamental interconnections between new product development and applications;
• understanding public perceptions of weather and improving public education accordingly;
• developing new ways to characterize forecast uncertainty and error and conveying the information in the most meaningful ways; and
• conveying information about forecast uncertainty and error to users.
Eve and Julie received more than 110 applications for 25 spots in the workshop. They were especially surprised at the range of disciplines represented in the applicant pool as well as the number of applications from senior agency officials and professors. Because of the great demand from scientists, graduate students, National Weather Service staff, and emergency managers and planners, they hope to find additional funding to support more WAS*IS workshops at NCAR and elsewhere. For more information about WAS*IS at NCAR.
Also in this issue...
UCAR family-friendly benefits rank at the top
New leadership for EOL
2005 Up-the-Hill Races
Random Profile: Vickie Johnson
NCAR's public face: Mesa Lab tour guides
Construction begins on CG-FL bike path
Hurricane Katrina Challenge
Just One Look
Staff Notes home page | News Center