Breaking records in the sky and sea: Innovations in research through the 2025 Atlantic hurricane season

The 2025 Atlantic hurricane season has officially come to an end. Throughout this record-breaking season, NOAA scientists conducted innovative research on tropical cyclones that will improve forecasting accuracy, enhance understanding of storm behavior, and strengthen preparedness efforts for communities. Their work and dedication contributed to a safer and more resilient future for those facing the impacts of extreme weather events.

As a part of NOAA’s Hurricane Field Program led by the Atlantic Oceanographic and Meteorological Laboratory (AOML), scientists gathered data from six different tropical cyclones during a total of 52 operational and research missions aboard the NOAA Hurricane Hunter aircraft. These missions started in August with Category 5 Hurricane Erin as it rapidly intensified in the Atlantic and continued through October as Category 5 Hurricane Melissa moved through the Caribbean.

Select research instruments used in the 2025 hurricane season: Hurricane Hunter aircraft, GPS dropsondes,  Black Swift drones, Oshen C-stars, underwater gliders.
Select research instruments used in the 2025 hurricane season: Hurricane Hunter aircraft, GPS dropsondes, Black Swift drones, Oshen C-stars, underwater gliders. Image credit: NOAA (Maria Raykova)

Hurricane research in the air

To gather crucial data from inside the storms, researchers at AOML flew with the NOAA Aircraft Operation Center crews aboard the Hurricane Hunter aircraft: two Lockheed WP-3D Orions (P-3s) and the high-altitude Gulfstream IV-SP (G-IV). Throughout a total of 52 flight missions, researchers released 1,112 GPS dropsondes and created 155 three-dimensional Tail Doppler Radar (TDR) wind analyses. Dropsondes are instruments that are released from the aircraft to collect pressure, temperature, humidity, and wind observations as they fall towards the ocean, while the TDR detects wind and precipitation within the hurricane. These instruments were critical for providing the National Hurricane Center (NHC) and Environmental Modeling Center (EMC) with real-time data that is used to improve forecast accuracy.

NOAA hurricane hunter flights gathered a number of record-breaking observations this season. One dropsonde in Hurricane Melissa observed a wind gust of 252 mph, breaking the record for the strongest wind ever measured by that instrument. During the same flight, a TDR wind analysis contained the strongest winds ever observed by the TDR in a hurricane. Other instruments on the plane also indicated record-breaking wind speeds during Melissa.

Researchers at AOML are also working with industry partners to pioneer advancements that revolutionize hurricane observing technology, ensuring more accurate forecasting and enhanced understanding of storm dynamics. Throughout this season, AOML successfully deployed 50 Skyfora StreamSondes, inexpensive, lightweight dropsondes that can be released in large numbers to extensively sample the low-level structure of several hurricanes. Scientists dropped the swarms of StreamSondes near the centers of the storms to understand the variations in their strongest regions.

NOAA also successfully launched 17 Black Swift TechnologiesS0 small uncrewed aircraft systems (sUAS). Observations from drones like the S0 allow scientists to gather data from previously inaccessible regions of the storm, laying the foundations for significant advancements in hurricane forecasting. During Hurricane Melissa, one Black Swift S0 had a record-breaking flight of 120 minutes and another became the first sUAS to capture video inside of a hurricane. In Hurricanes Gabrielle and Melissa, sUAS pilots successfully executed new flight maneuvers that proved that S0s can be positioned to target specific areas. These innovations will pave the way for the future of hurricane research technology.

First ever video captured by a Blackswift S0 sUAS inside of a hurricane. Video credit: Black Swift Technologies.

Hurricane research in the ocean

While the planes and atmospheric instruments were measuring the atmosphere, gliders, ocean drones, floats, and drifters were monitoring the ocean, which provides the energy for hurricane intensification. This season, NOAA teamed up with robotics company Oshen and the University of Southern Mississippi (USM) to deploy and operate 8 Uncrewed Surface Vehicles (USVs) called C-Stars to gather weather data at the ocean surface during the 2025 hurricane season. 

C-stars are four-foot-long, wind-propelled boats with solar-powered sensors that transmit real-time wind speed and direction, sea surface temperature, air temperature, air pressure, and relative humidity data every two minutes via satellite. They also take photos during their missions for immediate use, and video that is available after the vessels are recovered.

During Hurricane Humberto, one of the C-Stars became the first USV to gather data from the eyewall of a Category 5 hurricane. Two other C-Stars penetrated Hurricane Humberto to gather data while it was a Category 4 hurricane. 

NOAA, Oshen, and USM also rapidly deployed two C-Stars from the coast of North Carolina to collect data from Hurricane Imelda. The success of these rapid deployments demonstrated the benefit of using small, easily transportable USVs.

Photo of Oshen C-Star and big ocean wave, captured by an Oshen C-Star inside of Hurricane Humberto.
A close up image of a C-star in the water after deployment

(Photo 1) Image captured by a C-Star inside of Hurricane Humberto. Credit: NOAA/Oshen. (Photo 2) A close up image of a C-star in the water after deployment. Credit: Oshen

AOML and partners also operated 5 underwater hurricane glider missions in the Caribbean Sea and North Atlantic during the 2025 hurricane season to gather data from below the ocean’s surface. These gliders were part of a network monitoring the entire Atlantic basin collecting temperature and salinity data to depths of 900 meters. As a hurricane travels over the ocean, warmer surface water is mixed with cooler water below, reducing the energy available to fuel the passing storm. However, if there is a layer of low salinity water at the surface, it can serve as a cap that prevents mixing of ocean waters which can contribute to the intensification of tropical cyclones. The AOML gliders collected almost 6,000 temperature and salinity profiles over more than 470 cumulated glider-days, including data under 3 hurricanes: Erin, Imelda, and Melissa.

Hurricane Erin passed directly over an AOML glider, a US Navy glider, and a German glider, collecting observations that were cited by NHC forecasters who noted the intense cooling observed under the storm as a reason they were forecasting Erin to weaken. The P-3 aircraft coordinated with the gliders as part of NOAA’s Global Ocean Monitoring and Observing (GOMO) Program’s Extreme Events program to get a holistic view from the atmosphere down into the ocean. Ocean data can provide researchers and forecasters with a clearer understanding of ocean-atmosphere interactions, particularly how the ocean influences hurricanes.

Additionally, as Hurricane Melissa left the Bahamas, an AOML glider collected observations which showed a thick, warm ocean mixed layer at the surface extending down below. This deep reservoir of warm water provided a fuel source for Melissa, supportive of a possible re-intensification, which was noted by NHC forecasters.

An underwater hurricane glider just below the ocean’s surface. Credit: Cape Eleuthera Institute

AOML and partners also used argo floats and drifters to better understand the heat in the ocean that fuels hurricanes, and how it changes during the passage of storms, which is important for improving hurricane forecasts. Argo floats and drifters are deployed throughout the global ocean year round, however two Biogeochemical Argo floats were deployed alongside the hurricane gliders offshore Puerto Rico early in this hurricane season. 

During Hurricane Erin, NOAA’s Argo floats and drifters provided insights into the storm’s effects below the ocean’s surface. Argo profiles, processed by AOML, provided pre- and post-storm ocean conditions, confirming the low-salinity plume where Erin rapidly intensified. Drifters from the Global Drifter Program, along with air-deployed wave drifters, provided in-situ measurements, which were shared directly with NOAA’s NHC and Ocean Prediction Center. Additionally, during Hurricanes Gabrielle and Melissa, NOAA coordinated deployments of drifters and dropsondes near existing ocean glider paths as part of the Coordinated Hurricane Atmosphere-Ocean Sampling Experiment led by NOAA’s GOMO.

A yellow Argo float floating in the ocean.
An Argo float in the ocean after deployment.

Observations from various platforms across different depths of the ocean and atmosphere within a similar region gives researchers detailed data which is vital to advance forecast models, and improve hurricane prediction.

Innovations in modeling

This season, modelers at AOML spearheaded major advances in hurricane prediction by testing several next-generation configurations of the Hurricane Analysis and Forecast System (HAFS) in near real-time. Scientists were specifically working to upgrade the physics for HAFS-B and evaluate the new multi-storm versions of HAFS (HAFS-M and HAFS-C) which can track and simulate multiple tropical cyclones at the same time using high-resolution, storm-following nests.

As part of the Hurricane Forecast Improvement Program’s Real-Time Experimental effort, six experimental HAFS models—three from AOML and three from EMC—were run four times daily and compared to operational forecast models. These real-time tests provided critical insight into model behavior, tested different model configurations, and for the first time, demonstrated HAFS’s ability to extend forecasts beyond 5 days—advancing skill out to 7 days. In fact, throughout the 2025 season, the track forecasts from the experimental HAFS-M were 40% better than the operational Global Forecast System at Day 5 and nearly 50% better at Day 7. 

The experimental HAFS-B model was the best amongst all hurricane models at capturing the westward turn and extreme rapid intensification that Hurricane Melissa had south of Jamaica. Additionally, both the HAFS-B and HAFS-M models were better at predicting rapid intensification than the legacy hurricane model HWRF. The performance of the experimental HAFS this season highlights the strong potential of multi-storm, moving-nest technology for next year’s operations.


The dedicated scientists, flight crews, and staff that supported tropical cyclone research during the 2025 Atlantic Hurricane season spent months of long hours working diligently to gather data that will advance our understanding of storms and improve forecasts, with the ever present goal of protecting lives and property.

This work is accomplished in partnership with:

NOAA’s Aircraft Operations Center, Cooperative Institute for Marine and Atmospheric Studies, Environmental Modeling Center, Global Ocean Monitoring and Observing Program, National Data Buoy Center, National Environmental Satellite, Data, and Information Service, National Hurricane Center, Northern Gulf Institute, Ocean Exploration, Ocean Exploration Cooperative Institute, Office of Marine and Aviation Operations, Pacific Marine Environmental Laboratory, and Uncrewed Systems Operations Center, as well as the Office of Naval Research, National Air and Space Administration, Embry-Riddle Aeronautical University, University of Miami, University of Notre Dame, University of Southern Mississippi, University of the Virgin Islands, Stonybrook University, and Scripps Institution of Oceanography.