Jacobs now has the opportunity to build and strengthen programs that will help set the direction of weather prediction in the United States for decades to come.
He takes the job at a critical time. Last fall, our nation experienced one of its worst hurricane seasons on record, and the United States already has endured three billion-dollar weather disasters this year.
The primary U.S. weather forecasting system, often called the “American model,” has fallen short of the accuracy of competitors in Europe. While NOAA has invested substantial resources to improve the model, it has not been able to close the gap. Meanwhile, companies in the U.S. private sector are developing new tools and prediction systems from which the agency could benefit.
Leveraging his experiences at Panasonic, which housed a weather forecasting group to support the company’s aviation services, Jacobs has a great opportunity to help forge partnerships between the government and private sector so that U.S. weather forecasting can become the world’s best.
I recently sent some questions to Jacobs to inquire about what he hopes to accomplish at NOAA. His answers, lightly edited for length and format, are provided below.
Much has been made about the superiority of the global numerical weather prediction modeling systems in Europe vs. the U.S. Is the Trump administration taking any specific steps the narrow the gap and improve the U.S. global modeling system?
Accelerating the advancements in the U.S. global modeling program is a top priority of the administration and at NOAA. Numerical weather prediction is, at its core, an initial value problem, and to increase our predictive skill, we will be addressing three key areas: model code, observations, and computational resources. We need both better and more frequent observations, as well as the model code to assimilate these observations.
Earlier this year, NOAA upgraded its supercomputing system, making it among the fastest in the world, with the ability to process 8 quadrillion calculations per second, and adding 2.8 petaflops of speed to each data center, increasing NOAA’s total operational computing speed to 8.4 petaflops.
With this upgrade, U.S. weather supercomputing paves the way for NOAA’s National Weather Service to implement the Next Generation Global Prediction System, known as the “American Model,” which is being developed and tested in stages, and systematically transitioned into operations over the next two years.
In global modeling, satellite data are the driving force. The recently launched NOAA-20 [satellite] is providing global temperature and moisture information based on the best infrared and microwave sounder data.
When it comes to the model, we’ll be transitioning to a new dynamic core called the finite-volume cubed-sphere (FV3), which can be run at a much higher resolution. In addition to that, we will also be working on the model physics through the Common Community Physics Package.
We will be increasing both the horizontal (9 kilometers) and vertical resolution (128 levels) of both the deterministic model, as well as the global ensemble, which is a critical component of providing probabilistic guidance. This will be a more user-friendly “community” version of a global model, with the motivation that it will enable us to harness the collective enhancements and improvements from academic institutions and industry, thereby accelerating the research-to-operations “R2O” timeline.
Beyond improving numerical weather modeling capability a week or so into the future, what are some other weather prediction priorities? Can you describe efforts to improve forecasts between two weeks and two years into the future?
Seasonal to subseasonal forecasting is not just a priority, but part of the Weather Research and Forecasting Innovation Act which was signed by President Trump on April 18, 2017. Much of what we are learning in the global modeling area is also applicable to models that run further out. For example, two-way coupling between the atmospheric and ocean models, which is research we are conducting with other agencies, will be valuable for extending the predictive capabilities. Likewise, we will also be looking at other methods of analyzing longer-range signals in a statistical versus dynamical framework.
The 2017 Atlantic hurricane season was the most costly on record. What lessons can we take away from it in terms of building the nation’s capacity to prepare for and respond to future weather disasters?
The 2017 Atlantic hurricane season was a harsh reminder, especially after a record-setting 12-year period without a major hurricane making landfall on the U.S. mainland, that such powerful storms can and do hit the U.S., and they cause enduring harm to American lives and property.
Based on preliminary data, the National Hurricane Center’s Atlantic track predictions for 2017 set an all-time record for minimal position error across all forecast hours, which improved on the five-year mean error by about 25 percent.
These improvements, in addition to those made on the intensity forecast, are critical for extending the time window that decisions need to be made and allow for precise warnings to go out earlier to American communities, granting precious extra time to evacuate or find shelter. Likewise, employing a probabilistic approach from the global model ensembles allows us to quantify a level of confidence in the predictions.
The Weather Service employees’ union has expressed concern about forecasting vacancies at forecast offices but, at the same time, the NWS operations and workforce analysis conducted in 2016 concluded “that the current distribution of staff across the country can evolve.” What is the present thinking on how the structure and staffing of the Weather Service might evolve?
The Weather Service is a dynamic agency, with a workforce spread across the country at local forecast offices, national forecast centers and administrative headquarters. If approved by Congress, the Weather Service will begin implementing a series of operational reforms aimed at increasing staffing flexibility to best match service demands with available resources.
The leadership will accomplish this operational flexibility by reducing operational hours at various offices while maintaining meteorological services through collaboration with other Weather Service offices, making better use of collaborative forecast processes, technological innovation, and a more efficient forecaster career path, and moving away from a uniform staffing model to redistribute staff to best meet local partner needs.
NOAA looks forward to building a more constructive working relationship with National Weather Service Employees Organization [its labor union], as they are aligned in NOAA’s primary goal of continuing to provide America with exceptional science and service, while supporting the health and well-being of the workforce. Furthermore, any changes in the staffing structure will not affect NOAA’s commitment to providing local communities with forecast and warning coverage 24 hours a day, 7 days a week, 365 days a year.