Three and a half days before Hurricane Laura made landfall near Cameron, La., the National Hurricane Center predicted where it would come ashore within 0.6 miles. Not only did it peg the location near the Texas-Louisiana border but it forecast the exact hour it would cross the coastline: 2 a.m. Eastern.

In the days between that forecast and when the storm roared inland, the Hurricane Center barely wavered from its prediction, even as hundreds of computer model simulations forecast landfall locations as far-flung as Florida and Mexico.

Its precision forecast, made while Laura was still over Haiti, left meteorologists outside the agency in awe.

“The folks at the National Hurricane Center are so damn good at their job,” tweeted Dakota Smith, a meteorologist at the Cooperative Institute for Research in the Atmosphere.

“Just want to give a shout out to @NHC_Atlantic [the National Hurricane Center] and @NHC_Surge for the job they did w/ Laura. Lives were no doubt saved because of their efforts,” tweeted Matt Lanza, a meteorologist for SpaceCityWeather.com, a weather blog for the Houston area.

Even when the vaunted European computer model, the most accurate on average, suggested the storm would veer to the west and strike near Houston, the Hurricane Center didn’t flinch.

On Twitter, the Hurricane Center was the target of criticism and second-guessing when it didn’t shift its track forecast to place Houston and areas farther south under a hurricane warning.

“I saw a lot of comments online about not moving the track westward near Houston when this ECMWF [European model] ensemble plot came out,” tweeted Eric Blake, a forecaster at the Hurricane Center.

Blake said forecasters at the Hurricane Center knew the model “had been struggling this year (and with Laura) and adjusted accordingly.”

The Hurricane Center’s skilled and steady approach proved the doubters wrong. Blake’s message to them? “Trust the professionals and don’t just worship one model!”

Mike Brennan, chief of the center’s hurricane specialist unit, explained how it approaches such forecasts. “One of the things NHC prides itself on is the consistency of our track forecasts,” he said via email.

“We try not to follow short-term trends of any one model from one cycle to the next, but typically rely on a consensus or blend of the best 5 or 6 models as guidance for track forecasting. This helps smooth out the differences and biases between the models as well as the cycle to cycle variability in the model track forecasts. When we do make adjustments to the NHC track, we try and do so gradually, and wait to make large adjustments until we’re confident that a bigger shift in the models will hold over several model cycles.”

To many, the success with Laura affirmed the Hurricane Center’s status as the most trusted source for hurricane forecasts.

“This is why to rely on the @NHC_Atlantic tracks, and not those from some faceless, basement-dwelling, self-proclaimed ‘weather authority’ on social media,” tweeted Roger Edwards, a forecaster at the National Weather Service Storm Prediction Center.

“Nobody is perfect, but nobody is better, than the deeply experienced crew of world-class hurricane-prediction experts at NHC.”

As good as the Hurricane Center’s track forecast was for Laura, its predictions for its intensity at landfall weren’t as stellar.

Though it correctly forecast on Sunday morning the time and location of landfall, it predicted Laura would come ashore with peak winds of 100 mph, 50 mph too low, and had to play catch-up as the storm strengthened at breakneck speed.

The misfire on the intensity forecast illustrates a well-known gap between its storm track and intensity predictions.

Since the 1970s, the Hurricane Center’s average track errors have decreased by about two-thirds, but intensity forecasts showed very little improvement through the early 2000s.

Intensity forecasts have begun to improve over the past 15 years or so, however.

A study published this month in the journal Weather and Forecasting found a recent “notable improvement trend … that has been statistically significant and accelerating.”

Despite these gains, the study acknowledged that predicting when storms will rapidly intensify is a major challenge. Rapid intensification refers to at least a 35-mph increase in winds in 24 hours. Laura saw its winds increase by 65 mph in 24 hours, the fastest rate on record in the Gulf of Mexico, tied with one other storm.

“While we are better able to recognize the environmental conditions that support rapid intensification, knowing when it will start and how long it will last before an eyewall replacement cycle occurs or some other factor stops the intensification trend is still quite difficult,” wrote Brennan.

Better understanding rapid intensification is a major focus of hurricane research, and new prediction techniques are being developed.

“Much work still needs to be done to further improve [rapid intensification] forecasting, and as new and existing tools are developed and refined, there should be an increasing confidence in applying RI forecasts in operations,” the Weather and Forecasting study said.