New Methodology Improves Winter Climate Forecasting

It’s hot out right now, but new research from North Carolina State University will help us know what to expect when the weather turns cold. Researchers have developed a new methodology that improves the accuracy of winter precipitation and temperature forecasts. The tool should be valuable for government and utility officials, since it provides key information for use in predicting energy consumption and water availability.

“Predicting winter precipitation is extremely useful, because winter is the most important season in terms of re-charging water supplies in the United States, ensuring water will be available in the summer,” says Dr. Sankar Arumugam, author of the study and an assistant professor of civil, construction and environmental engineering at NC State. The study was co-authored by Naresh Devineni, a Ph.D. student at NC State.

“Predicting temperature is also important, because temperature determines energy consumption,” Arumugam says. “When it is very cold, people use more energy to heat their homes.”

The researchers were able to reduce uncertainty in winter climate predictions over the United States by developing a methodology that incorporates multiple general climate forecast models (GCMs) and also accounts for the activity – or inactivity – of El Nino conditions in the Pacific.

Winter precipitation and temperature over many regions of the continental United States are predominantly determined by the El Nino Southern Oscillation (ENSO), which denote hot (El Nino) or cold (La Nina) sea surface temperature conditions in the tropical Pacific.

Most  GCMs are better at predicting the winter climate when ENSO is quite active, and are less accurate under neutral ENSO conditions. The methodology developed by the researchers accounts for the skill of the models under active and neutral ENSO conditions in combining multiple GCMs, resulting in reduced uncertainty in predicting the winter climate.

“Improving precipitation and temperature predictions should help government, water and energy utility officials plan more effectively,” Arumugam says, “because they will have a better idea of what conditions to expect.”

The study, “Improving the Prediction of Winter Precipitation and Temperature over the continental United States: Role of ENSO State in Developing Multimodel Combinations,” was published online this month by Monthly Weather Review. The research was funded by the North Carolina Water Resources Research Institute.

NC State’s Department of Civil, Construction and Environmental Engineering is part of the university’s College of Engineering.

-shipman-

Note to editors: The study abstract follows.

“Improving the Prediction of Winter Precipitation and Temperature over the continental United States: Role of ENSO State in Developing Multimodel Combinations”

Authors: Naresh Devineni, A. Sankarasubramanian, North Carolina State University

Published: June 2010 (made available in July), Monthly Weather Review

Abstract: Recent research in seasonal climate prediction has focused on combining multiple atmospheric General Circulation Models (GCMs) to develop multimodel ensembles. A new approach to combine multiple GCMs is proposed by analyzing the skill of candidate models contingent on the relevant predictor(s) state. To demonstrate this approach, we combine historical simulations of winter (December-February, DJF) precipitation and temperature from seven GCMs by evaluating their skill – represented by Mean Square Error (MSE) – over similar predictor (DJF Nino3.4) conditions. The MSE estimates are converted into weights for each GCM for developing multimodel tercile probabilities. A total of six multimodel schemes are considered that includes combinations based on pooling of ensembles as well as based on the long-term skill of the models. To ensure the improved skill exhibited by the multimodel scheme is statistically significant, we perform rigorous hypothesis tests comparing the skill of multimodels with individual models’ skill. The multimodel combination contingent on Nino3.4 show improved skill particularly for regions whose winter precipitation and temperature exhibit significant correlation with Nino3.4. Analyses of weights also show that the proposed multimodel combination methodology assigns higher weights for GCMs and lesser weights for climatology during El Nino and La Nina conditions. On the other hand, due to the limited skill of GCMs during neutral conditions over the tropical Pacific, the methodology assigns higher weights for climatology resulting in improved skill from the multimodel combinations. Thus, analyzing GCMs’ skill contingent on the relevant predictor state provide an alternate approach for multimodel combination such that years with limited skill could be replaced with climatology.