Weather Forecasts and Why the Models Matter
Mel Kent
Northern Virginia Community College
Author Note
This paper is being submitted on XXX,
for Professor Bartlett’s ITE 119 Section 001W.
Outline
I.
Introduction
A.
Some forecasting models absolutely butcher the
prediction of East Coast snowstorms.
B.
Tech and big data work together to bring us
up-to-date weather forecasting models, but not all models are created equal.
The European Center for Medium-Range Weather Forecasts (EURO) does a
significantly better job predicting major East Coast winter storms than the
Global Forecast System (GFS).
II.
GFS Model has a bad habit of overplaying the jet
stream and underplaying low pressure systems, as well as misreading temperature
anomalies in the lower atmosphere, VS high-computing Euro sees them better.
A.
Previous winter storms inaccuracy in
forecasting.
1. Winter Storm Jonas in 2016.
( https://search-proquest-com.eznvcc.vccs.edu/globalnews/docview/2344542552/17B86BFB666D4738PQ/3?accountid=12902
)
2.Winter Storm Petra in 2019 ( https://search-proquest-com.eznvcc.vccs.edu/globalnews/docview/2186333243/C934203043DE4B90PQ/5?accountid=12902
).
B.
December 2020 storm comparison. ( https://search-proquest-com.eznvcc.vccs.edu/globalnews/docview/2469557523/2A21E00E02AE44DCPQ/4?accountid=12902
)
C.
GFS Ensemble, transition to Euro ensemble.
III.
The EURO model runs an ensemble of different
models and averages them together. Takes more computing power, only run twice a
day. (https://search-proquest-com.eznvcc.vccs.edu/globalnews/docview/2287029851/93790D1A4FAC4C3CPQ/1?accountid=12902
)
A.
The European community charges up to $250k for
all of their info, so they have more $$ for research and development.
1. “Initialization” determining
the current weather conditions used as a starting point to feed the models.
2. Euro “The King” UKMET “The
Queen.”
B. 200 Gigabit HDR InfiniBand selected to
accelerate their new supercomputer.
C.
Supercomputers and tech updates to increase
accuracy in weather forecasting. Numerical predicting activities. Supercomputer
designed to maximize energy efficiency. ( https://search-proquest-com.eznvcc.vccs.edu/globalnews/docview/2336854995/CF21D572405D4411PQ/3?accountid=12902
)
IV.
GFS needs to be retired, to be replaced by the
FV3 (Finite-Volume Cubed-Sphere dynamical core”. (https://search-proquest-com.eznvcc.vccs.edu/globalnews/docview/2180372588/3457E68275DC400DPQ/13?accountid=12902
)
A.
FV3 has been tested through 3 summers, hurricane
seasons, and winters. Mets so far are not fans, as it is just about as
inaccurate as the GFS. Is it worth it to grow with the system?
1. Grow and adapt with the new
program, learn to see the biases and work around them.
2. Agreement between NOAA and National
Center for Atmospheric Research (NCAR, nonprofit, federally funded) for more
collaboration between forecasters and researchers to help improve models.
B.
Earth Prediction Innovation Center (EPIC) – a
cloud-based clearinghouse for researchers developing new now and emerging model
tech. Crowdsourcing with tech to find better solutions. (https://search-proquest-com.eznvcc.vccs.edu/globalnews/docview/2280205829/93790D1A4FAC4C3CPQ/2?accountid=12902
)
C.
Possibility for GFS to overcome negative bias
through this new program.
V.
Conclusion
A.
GFS is shaky on catching weather events because
low computing power. Euro has higher computer power, making it more accurate.
Euro and GFS are both being updated, possibility for GFS to regain some of it’s
positive notoriety. Tech innovations and data sharing, weather through EPIC or
otherwise, will help improve forecasting capabilities.
B.
In the future, we may have better access to
higher quality prediction software, which will make our lives on the densely
populated East Coast easier to plan during winter. Better computers = better
models = better us!
