Groups Similar Look up By Text Browse About



Similar articles
Article Id Title Prob Score Similar Compare
139292 ARSTECHNICA 2019-6-12:
The long-awaited upgrade to the US weather forecast model is here
1.000 Find similar Compare side-by-side
139322 ENGADGET 2019-6-12:
US weather forecast model gets a much-needed upgrade
0.990 0.746 Find similar Compare side-by-side
139040 VENTUREBEAT 2019-6-11:
Amazon researchers bolster Alexa’s language learning with preexisting knowledge
0.002 0.437 Find similar Compare side-by-side
139694 VENTUREBEAT 2019-6-14:
Google’s Evolved Transformer achieves state-of-the-art performance in translation tasks
0.001 0.404 Find similar Compare side-by-side
138884 VENTUREBEAT 2019-6-10:
Facebook launches PyTorch Hub for reproducing AI model results
0.357 Find similar Compare side-by-side
139520 VENTUREBEAT 2019-6-13:
AI improves Alexa’s error rate with challenging training sets
0.345 Find similar Compare side-by-side
139077 VENTUREBEAT 2019-6-11:
Intel researchers compress AI models without compromising accuracy
0.338 Find similar Compare side-by-side
139665 TECHREPUBLIC 2019-6-14:
Linux-powered System76 Gazelle laptop allows users to avoid the Windows Tax
0.329 Find similar Compare side-by-side
139023 ARSTECHNICA 2019-6-11:
It watches you drive: Subaru Forester review
0.324 Find similar Compare side-by-side
139187 TECHCRUNCH 2019-6-11:
Amazon Alexa team uses machine learning to better handle regional language differences
0.320 Find similar Compare side-by-side
139754 ENGADGET 2019-6-15:
Tesla starts selling used Model 3 cars online
0.307 Find similar Compare side-by-side
139448 THENEXTWEB 2019-6-13:
Apple just registered 7 laptop models we might see this fall
0.295 Find similar Compare side-by-side
139043 ENGADGET 2019-6-11:
Logitech updates a trio of gaming mice with high-precision sensors
0.291 Find similar Compare side-by-side
139378 VENTUREBEAT 2019-6-12:
Essential tips for scaling quality AI data labeling
0.270 Find similar Compare side-by-side
138952 VENTUREBEAT 2019-6-10:
AI helps drones dodge fast-moving objects
0.264 Find similar Compare side-by-side
139523 TECHCRUNCH 2019-6-13:
Expedition installs highest weather stations on Earth atop Mount Everest
0.254 Find similar Compare side-by-side
139600 VENTUREBEAT 2019-6-14:
AI Weekly: ICML 2019 top papers and highlights
0.254 Find similar Compare side-by-side
139455 VENTUREBEAT 2019-6-13:
Ditch the data scientists and weaponize your data with AI tech (VB Live)
0.251 Find similar Compare side-by-side
139316 THEVERGE 2019-6-12:
LG’s 55-inch C7 OLED 4K HDR TV is under $1,000 for the first time
0.249 Find similar Compare side-by-side
139215 THENEXTWEB 2019-6-12:
NASA research shows alien life is far less likely than previously thought
0.240 Find similar Compare side-by-side
139259 THEVERGE 2019-6-12:
Nintendo starts producing new Switch models outside China to fend off trade war
0.240 Find similar Compare side-by-side
138983 VENTUREBEAT 2019-6-11:
IBM’s Watson Studio AutoAI automates enterprise AI model development
0.239 Find similar Compare side-by-side
138974 VENTUREBEAT 2019-6-10:
Google launches TensorFlow.Text library for language AI models
0.237 Find similar Compare side-by-side
139018 VENTUREBEAT 2019-6-11:
The 3 critical AI research questions (VB Live)
0.236 Find similar Compare side-by-side
139643 ARSTECHNICA 2019-6-14:
Study finds that a GPS outage would cost $1 billion per day
0.236 Find similar Compare side-by-side

1

ID: 139292

URL: https://arstechnica.com/science/2019/06/us-weather-forecasts-get-a-software-update/

Date: 2019-06-12

The long-awaited upgrade to the US weather forecast model is here

Its been almost 40 years since the model got a new core. Weather forecasters need a ton of knowledge and a fair bit of experience with local weather patterns to do their job well. They also need a good forecast model. These computer models take in measurements from weather stations on the ground, satellites in orbit, and balloons in between and then simulate the physics of weather forward in time a few days. For the first time in about 40 years, the guts of the US model got swapped out for something new today.  The upgrade brings us a new Finite-Volume Cubed-Sphere (or FV3) dynamical core, which simulates the basic atmospheric physics at the heart of this endeavor, a change that has been in the works for a while. The new core had its origins in simulating atmospheric chemistry but ended up being adapted into other models. A few years ago, it was selected to replace the old core in the US Global Forecast System model. And for more than a year now, the new version of the model has been running in parallel so its results could be compared to the operational model. That evaluation also included retroactive forecasts (that is, forecast simulations with the same inputs that were available on each day in the past) of the past three years, with an additional focus on case studies of major hurricanes and common storm types. The results have been a little mixed. The new core improves computational efficiency and allows some processes to be simulated at a higher resolution—unequivocal improvements. It also simulates the physics of water vapor more realistically. In a press conference today, NOAA scientists cited a number of areas where forecast improvements have been seen. Forecast tracks of hurricanes and the mid-latitude storms that frequently sweep across the US have both improved, they said, along with forecasts of hurricane strength. Forecast precipitation amounts were also cited as a key area of progress. But there have also been grumblings in the weather community over the past year about results that didnt seem so hot. For example, surface temperatures have been biased low in some situations, throwing off forecasts. Of course, no model is a perfect replica of the planet, and part of using these models to forecast weather requires us to understand and account for their biases. NOAA says that the cold bias has been reduced by development over the past few months—and development is a continuous process that doesnt stop when a model goes live. Still, there could be an adjustment period here as forecasters get used to their new souped-up model. At least through September of this year, the old model will still be running in parallel to aid in that transition. Improvements to the forecast system can come from other fronts, too, from recent upgrades to NOAAs supercomputers that drive the models to launches of the latest and greatest satellites. But switching out the core of the main forecast model is a big deal that will hopefully help it keep up with the progress being made by others. Right now, the European Centre for Medium-Range Weather Forecasts model is generally seen as the top dog.