3 min readCompany news

The FOX Model: a broadcast weather model developed with TempoQuest

In July 2023 TempoQuest announced that FOX Weather was working with it to run the FOX Model, four times a day, out to 84 hours. Here is what was said and what a model like that takes.

A FOX Weather broadcast map of current temperatures across the United States, ranging from 50 degrees in Rapid City to 92 degrees in Phoenix.

First published July 31, 2023 · Updated November 16, 2026

A broadcast meteorologist works to a hard deadline. The forecast has to be ready before the show, and it has to show the viewer's own county, not a smooth national picture. Public models run on fixed schedules over fixed areas, and every station has the same ones. Some broadcasters answer that by running a model of their own. In July 2023, TempoQuest announced that FOX Weather was working with TempoQuest to run one: the FOX Model.

What was announced

TempoQuest published its announcement on July 31, 2023. FOX Weather described the model in its own article, first published in December 2021. Taken together, they said:

  • The FOX Model was an exclusive high-resolution forecast model for FOX Weather.
  • It was first developed by meteorologist Jim Weber at FOX 13 Tampa with TempoQuest.
  • It ran four times a day.
  • Each run forecast out to 84 hours.

FOX Weather's article presented the model first as a tool for tracking tropical systems, and said its coverage area could be moved to other severe weather around the country.

These details are as announced at the time.

Why a broadcaster runs its own model

Three things change when a station or a network runs its own regional model.

  • The area. A regional model covers a chosen box at fine grid spacing instead of the whole globe at coarse spacing. The box can sit where the story is: a coastline ahead of a hurricane, or the Plains during a tornado outbreak.
  • The clock. Each new run starts from newer observations. Four runs a day means a fresh forecast about every six hours.
  • The picture. The team controls the output and the graphics, so the forecast on screen is its own and not a map that every competitor also has.

A forecast out to 84 hours covers three and a half days. For a landfalling hurricane, that spans the window in which people decide whether to board up, stock up or leave.

What it takes to run a model four times a day

A six-hour cycle sets a hard budget. Each run has to finish, with time left to make graphics, before the next one starts. Resolution is what strains that budget.

Finer grids need much more computing.

That arithmetic is why hardware matters to a broadcast model. AceCAST, TempoQuest's product, runs WRF, the Weather Research and Forecasting model maintained by NCAR, on NVIDIA GPUs. It keeps the WRF workflow: WRF-compatible inputs, namelists and outputs.

See the benchmark and cost example.

If your team needs something similar today

Not every organization wants to operate a model. There are three routes, depending on how much of the work you want to own.

  • Run the model yourself. AceCAST runs on Linux systems with NVIDIA GPUs, on premises or in the cloud. A trial runs for 14 days on one machine.
  • Order a forecast for a site. Aerava, TempoQuest's weather intelligence platform, runs storm forecasts at 1 km over a 500 km square in the US, up to 12 hours ahead. The price is shown before you run. Aerava states that its storm forecasts are not yet scored against observations.
  • Watch live weather on screen. WSV3 is TempoQuest's proprietary weather visualization platform for live radar, GOES-R satellite imagery, lightning, warnings and forecast data. TV broadcasting is one of the fields it is used in.

See AceCAST 5, available now, for its physics, outputs and upgrade guidance.

What to do next

Request a trial, or talk to us about a forecast model for your newsroom.

Your forecast

Have a forecast question of your own?

Tell us the decision it has to serve. One short form reaches the team, and we review every request.