How a 3-Year Balloon Project Became NASA's 25-Year Secret to Tracking Global Ozone

For climate scientists and meteorologists tracking the health of Earth's atmosphere, the NASA SHADOZ network has become an indispensable tool for calibrating satellite ozone measurements. Originally proposed in 1998 as a modest three-year project to fix a tropical blind spot in satellite algorithms, the Southern Hemisphere ADditional OZonesondes (SHADOZ) network has now surpassed 25 years of operation. By October 2023, the project had successfully archived 10,000 ozone and pressure-temperature-humidity profile pairs.

This milestone provides researchers with a critical baseline to verify orbital data from missions like the Aura satellite, ensuring that global climate models remain accurate. Without these physical balloon measurements, retrieval algorithms would struggle to interpret the vertical distribution of ozone across a region equivalent to about 35 to 40 percent of Earth's surface. Today, the archive receives more than 300,000 data-user hits annually, with over 20 publications referencing SHADOZ data in 2023 alone.

An ozonesonde is a lightweight instrument carried beneath a standard rubber weather balloon alongside a meteorological radiosonde. According to NOAA's Global Monitoring Laboratory, the device uses an electrochemical sensor that pumps ambient air through a potassium iodide solution. As the air passes through, the ozone produces an electrical signal proportional to the amount present, which the radiosonde then transmits back to a ground station along with pressure, temperature, and humidity readings.

A typical balloon climbs to roughly 35 kilometres in about two hours before bursting, capturing a highly detailed vertical profile from the surface through much of the stratospheric ozone layer. While these instruments are generally treated as expendable because recovery is uncertain, they are not always lost permanently. NOAA's Hilo ozonesonde program in Hawaii, for example, reports successfully recovering and reusing a minority of its instruments after they descend back to Earth.

During the 1970s and 1980s, ozone observations were heavily concentrated at mid and high latitudes, leaving the tropics severely undersampled. Natal, Brazil, supplied the main regular tropical soundings during this era, supplemented by intermittent measurements from American Samoa and Hawaii. This gap was problematic because satellite instruments cannot simply read a perfectly resolved ozone profile from orbit; they require direct physical profiles to develop and check their retrieval algorithms.

Major field campaigns eventually built the infrastructure needed to close this gap. NASA's TRACE-A campaign in 1992 utilized aircraft and ozonesondes across the tropical Atlantic, while the PEM-Tropics-A mission in 1996 extended these measurements across the Pacific. By the time Anne Thompson and her colleagues at NASA Goddard organized SHADOZ in 1998, the necessary facilities, launch gas, and trained operators were already in place to standardize the data and build a common archive.

SHADOZ survived far beyond its initial three-year mandate because a coordinated tropical record became exponentially more useful as it grew longer. By its 2023 milestone, the partnership involved organizations from 14 nations across five continents, producing roughly 20 percent of the data from long-term ozonesonde stations worldwide. NASA's 2023 station inventory identified 14 stations with at least 10 years of operation, and despite historical interruptions at sites like Tahiti, Malindi, Cotonou, and Watukosek in Java, the 2023 archive still accumulated about 400 profile pairs from its long-term stations.

Maintaining accuracy across so many international sites requires rigorous quality control. The 2017 Jülich Ozonesonde Intercomparison Experiment tested these standards, and a resulting 2019 paper in the Bulletin of the American Meteorological Society reported that SHADOZ stations following recommended protocols measured total ozone within about 3 percent of the Jülich reference instrument. This two-way check ensures that sondes remain crucial for satellite validation, while mature satellite records can expose any drifts in a station's balloon data.

The true value of a multidecade archive emerges when researchers analyze long-term trends. A 2021 study in the Journal of Geophysical Research: Atmospheres examined SHADOZ profiles from 1998 through 2019 across five tropical regions, finding that free-tropospheric ozone generally showed its strongest positive trends from February through May. Adding three more years of observations through 2022 did not overturn these interpretations in NASA's 2023 assessment.

A newer 2025 analysis in Atmospheric Chemistry and Physics extended the trend calculations through 2023. The authors reported that the additional four years changed the picture very little: total tropospheric ozone trends were generally small, around 0.5 to 1 Dobson Unit per decade, except over Southeast Asia. Furthermore, the sonde data and OMI/MLS satellite trends agreed within their uncertainties at four of the five analysed sites. Today, the current NASA Goddard SHADOZ archive lists 16 active sites, including Quito and Palau, and continues to recommend its Version 6 data product, while Version 7 is planned as an eventual reprocessing effort.

The enduring success of the SHADOZ network highlights a fascinating reality in modern Earth observation: multi-billion dollar orbital satellites still fundamentally rely on rubber balloons and chemical cells to verify their accuracy. While remote sensing technology has advanced exponentially since 1998, the physical ground-truth sampling provided by an ozonesonde remains irreplaceable for calibrating complex retrieval algorithms.

The 2025 analysis proved this point by showing that adding thousands of commercial-aircraft profiles produced little change in the derived ozone trends. This indicates that SHADOZ's relatively sparse but highly consistent balloon schedule is more than capable of resolving multidecade atmospheric behavior. Ultimately, the network proves that in climate science, long-term methodological standardization is often far more valuable than the sheer volume of data collected.