Friday, October 20, 2017

Since Katrina: A decade of NOAA hurricane research

Thursday, August 27, 2015

Advancing hurricane forecasting, improving models, and increasing observations

by Caroline Mosley (NOAA Research Communications)

A decade ago, the United States experienced one of the most active and destructive hurricane seasons ever recorded. The loss of life and destruction of property from Hurricanes Katrina (Aug. 29), Rita (Sept. 24), and Wilma (Oct. 24) drove NOAA to re-evaluate hurricane research and severe storm preparedness.

Research and collaboration at NOAA’s Oceanic and Atmospheric Research (OAR) focuses on advancing our understanding of the science behind hurricanes and applying that knowledge to improve hurricane modeling. OAR provides the critical research and sustained modeling efforts required to improve hurricane forecasts, with the eventual goal of transitioning models to operational use for NOAA’s National Weather Service (NWS) to increase forecast accuracy and improve public preparedness and response to storms.

“Since 2005, NOAA has enhanced observations, upgraded to more powerful supercomputers, and improved models to better represent hurricanes,” said OAR Assistant Administrator Craig McLean, “With a continued level of focus, we can better understand hurricanes as part of a global system and enhance both our storm-specific and seasonal hurricane models.”

Advancing Hurricane Forecasting

Evolution of Hurricane Katrina with HWRF

Evolution of Hurricane Katrina with HWRF

This is a detailed view of the evolution of Hurricane Katrina when HWRF runs at a resolution of 2 km. This model can show small structures inside the hurricane, which is critical for providing better forecasts. Credit: NOAA/NWS

In 2008, Congressional support allowed OAR scientists and forecasters to begin a concentrated effort on the Hurricane Forecast Improvement Project (HFIP) which continues today. 

“After Hurricane Katrina, NOAA’s hurricane program HFIP emerged to identify the best ways to accelerate the improvement of hurricane forecasts and increase community resiliency,” says Robert Atlas, Ph.D., Director of NOAA’s Atlantic and Oceanographic Meteorological Laboratory (AOML).

Since HFIP, significant improvements such as increasing model resolution and reducing the forecast error help NWS provide more accurate information to the public. Increased partnerships encourage scientists and forecasters to work side-by-side, so scientist better understand and develop products that meet the needs of forecasters.

In the last five years, HFIP has reduced track and intensity forecast errors by twenty percent and upgraded the Hurricane Weather Research and Forecasting (HWRF) model so the finest resolution went from 9 to 2 kilometers. In 2015, the HWRF model was expanded to simultaneously simulate multiple storms in all hurricane basins globally, resulting in further forecast improvements.

Improving Models

In past years, research enhanced the quality of hurricane track forecasts, while improvements to hurricane intensity forecasts lagged behind. New research and development since Hurricane Katrina using the HWRF and Geophysical Fluid Dynamics Laboratory (GFDL) operational hurricane forecasting models helped to reduce errors in simulated hurricane intensity.

“The GFDL regional hurricane model has been operational through the National Weather Service since 1995,” said Morris Bender, a senior researcher at GFDL on the hurricane research team since 1976. Morris added that the sustained focus at GFDL has allowed NOAA to continually improve predictions of hurricane intensity.

The physical representation of hurricanes in models also leads to even better hurricane intensity and track forecasts. Work on HWRF at NOAA’s Atlantic and Oceanic and Meteorological Laboratory (AOML) and the Earth System Research Laboratory (ESRL) focuses on improving these physical representations.

GFDL HiFLOR Prototype Seasonal Prediction Model

GFDL HiFLOR Prototype Seasonal Prediction Model

This video simulates global hurricane activity and year-to-year variations in the seasonal number of intense hurricanes in the Atlantic from August through December of one year. Hurricanes are highlighted by dashed circles on the map while the time series records the evolution of each individual instantaneous hurricane’s simultaneous wind speed. Credit: NOAA/GFDL

Increased resolution in the HWRF, as well as GFDL models, helps to better represent small-scale hurricane structures that drive the evolution of the storms. Aided by aircraft observations collected in hurricanes, the models also include a better physical representation of the atmosphere and ocean as well as an improved simulation of small-scale interactions between these two regions.

One example of this since Hurricane Katrina is the incorporation of the Loop Current, a deep warm current in the Gulf of Mexico, into the operational GFDL regional hurricane model. GFDL scientists collaborated with the University of Rhode Island to incorporate this feature, which contributed to the rapid strengthening of Hurricane Katrina in 2005. The inclusion of the Loop Current helps to better represent how warm ocean waters, both at the surface and at depth, can elevate storm intensity.

GFDL also developed a high-resolution prototype seasonal prediction model (HiFLOR) that can realistically simulate global hurricane activity and predict the year-to-year variations in the seasonal number of intense hurricanes in the Atlantic. This experimental model targets extreme tropical cyclones, especially category 4 and 5 hurricanes. 

"This system promises to provide valuable information for improved seasonal outlooks of hurricane activity and has already provided model-based forecast guidance to the National Weather Service," explains GFDL researcher Gabriel Vecchi.

Since these hurricanes are responsible for nearly eighty-five percent of hurricane damage, improved forecasting of hurricane season characteristics months to seasons in advance could provide a better outlook of what to expect in the future.

Increasing Observations

Even with these advancements in hurricane modeling and forecasting, OAR continually seeks to improve weather models. Satellites from the National Environmental Satellite, Data, and Information Service (NESDIS) provide a view of hurricanes from a global scope, while increased weather observations provide field measurements for models.

NOAA P3 and GI-V aircraft

NOAA P3 and GI-V aircraft

These NOAA P3 and GI-V aircraft are used to collect data from directly inside hurricanes. Credit: NOAA/AOC

Starting in 2005, the AOML Hurricane Research Division (HRD) began conducting the NOAA Intensity Forecast Experiment (IFEX), developed through a partnership with NWS and NESDIS. IFEX missions use the two NOAA P-3 and G-IV aircraft operated by NOAA's Aircraft Operations Center (AOC) to collect and provide hurricane data for research and applications. Since Hurricane Katrina, a main goal of IFEX is the collection of hurricane data to directly aid the development and evaluation of HWRF.

NOAA leads missions over the Atlantic Ocean on the NASA Global Hawk unmanned aircraft to collect real-time data that can inform NOAA’s National Hurricane Center (NHC) forecasts. OAR is also testing the use of another unmanned aircraft to collect and report real-time observations in the area where the ocean meets the atmosphere. P-3 Hurricane Hunter aircrafts deploy the Coyote Unmanned Aircraft System (UAS) which drops down to collect and report real-time observations in this area below 3,000 feet, where manned aircraft cannot safely fly. 




Latest Imagery

Research Videos

Connect with

Office of Oceanic & Atmospheric Research Headquarters

1315 East-West Highway | Silver Spring, MD 20910 | 301-713-2458