| The Engineering Research
Center (ERC) for Collaborative Adaptive Sensing of the Atmosphere (CASA)
deploys short-range Doppler radar networks in the lowest level of the atmosphere
to overcome predictive limitations of conventional radar—an approach validated
by National Weather Service forecasters who reviewed data from the Center’s
systems and found it would have made their past assessments more accurate.
The NSF-funded research center is based at the University of Massachusetts,
Amherst.
In fall 2008, 16 active National
Weather Service (NWS) forecasters reviewed six different archived weather
cases to measure the impact of CASA’s high-resolution, lower troposphere
data on their wind predictions. The forecasters reviewed cases with data
gathered only from the current NEXRAD radar technology, and cases with
data from the current technology and from the Center’s networks of distributed,
collaborative sensing systems working in the lower atmosphere. The
NWS forecasters compared their predictions to measurements from ground-based
sensors.
The results show that forecasters
using data from both the current technology and CASA’s network’s made wind
assessments that were 30 percent more accurate, as measured by mean absolute
error, than when assessments were made only with data from the current
NEXRAD technology—a statistically significant improvement. Furthermore,
using a seven-point scale, forecasters also self-reported being significantly
more confident in their assessments when using data from the research center’s
lower-atmosphere radar networks.
Based on these results, it
is likely that these more accurate and confident wind assessments will
lead to improved severe thunderstorm warnings, which the NWS issues to
alert the public when it forecasts wind speeds in excess of 50 knots (58
mph). Forecasters said that with access to data from CASA’s networks,
they would issue warnings more often for the cases where actual wind speeds
were close to or above the severe storm threshold. Equally important,
forecasters chose not to issue warnings for cases where actual wind speeds
were under severe limits. This impact on warnings will be verified
in subsequent studies. |
