We need more, not fewer, measurements - and the tools to make them actionable

March 23 was World Meteorological Day, following World Water Day on March 22, and World Glaciers Day on March 21. A lot of days relating to the hydro-cryo-atmosphere all in a row, with World Meteorological Day commemorating the establishment of the World Meteorological Organization (WMO) in 1950. Happy 75th to the WMO, and many more to come!

March 23 was also the end of a weekend that found me, and many of my meteorology colleagues, thinking about, discussing, and answering questions related to the loss of meteorological data. Since February 27, the National Weather Service has announced the suspension of weather balloon flights at Kotzebue, Alaska; Omaha, Nebraska, and Rapid City, South Dakota; intermittent suspensions at Albany, New York, and Gray, Maine; and reductions of weather balloon flights at Aberdeen, South Dakota, Gaylord, Michigan, Grand Junction, Colorado, Green Bay, Wisconsin, North Platte, Nebraska, and Riverton, Wyoming, because of staffing shortages. Adding in the suspension of radiosonde launches at Denver, Colorado, in 2022 and Tallahassee, Florida, in 2024 because of the global helium shortage, and the decommissioning of the Chatham, Massachusetts, site in 2021 because of coastal erosion, 14 of the approximately 100 radiosonde sites that the National Weather Service operates in the United States, South Pacific, and Caribbean, have stopped flying balloons. [Weather balloon and radiosonde will be used interchangeably in this article.] I shared a map and some initial thoughts about the reduction of flights on my LinkedIn profile on Thursday.

I've spent more than a decade in the weather measurements space and was using assorted weather measurements for another decade before that, so I have a pretty good handle on the measurements that are collected globally and what we lose without them. I've shared some thoughts on the subject with local TV stations in Green Bay and Madison, Wisconsin, and NOTUS, but I am writing this piece to expand beyond what was included in the previous interviews and to consolidate my thoughts on weather measurements in one location.

Weather Balloons and Radiosondes

For much of my career, I worked for Vaisala , one of the companies which manufactures the radiosondes which fly on weather balloons. Weather balloons have been flying for more than 100 years and collect measurements from the surface of Earth all the way to the stratosphere, at heights of more than 35 km (roughly 115,000 feet).

We need radiosondes.

All the weather that happens at the surface starts above us somewhere else. It would be impractical to have 100,000-foot-tall weather stations around the world from safety, financial, and logistical standpoints, so we attach lightweight all-in-one weather stations (radiosondes) to large balloons and release them into the sky. These flights occur at around 900 locations around the world and are supposed to occur twice per day - once at 0000 UTC and once at 1200 UTC - giving twice-daily snapshots of the state of the whole atmosphere.

Radiosondes measure temperature, humidity, wind speed and direction, and air pressure at approximately 5-meter vertical resolution as they ascend through the atmosphere. There is no other sensing technology that can provide in-situ measurements through the atmospheric profile like a radiosonde. Profiling LIDARs, wind profilers, and satellite retrievals can give some of the data a radiosonde can, but at lower resolution and higher cost, and their data are truthed by radiosonde data.

In fact, over the previous year (March 23, 2024 - March 22, 2025), radiosondes were the second-most impactful measurement to the GEOS-Forward Processing data assimilation system. And this impact was felt with fewer observations than the most impactful measurement.

Losing these observations will hurt short- and long-term forecasting. Weather computer models rely on data from radiosondes to inform the initial conditions that all model forecasts start with. Missing large sections of real data means that the models must first interpolate the current conditions (creating errors) and then calculate forward from there, creating even larger errors. This is an oversimplification of chaos theory in weather modeling, but Edward Lorenz, Ellen Fetter, and Margaret Hamilton showed in the 1960s how small errors in the initial conditions lead to large errors the further in time you go.

Critically, radiosondes provide important measurements for short-term, high-impact weather that other measurement platforms can't. Because of their high-resolution (5-meter) measurements, radiosondes can detect subtle changes in the vertical profiles of temperature, humidity, and wind. It is often these small changes that have big impacts. A small temperature inversion (temperature rising, instead of falling, with height) can trap pollutants or wildfire smoke, leading to unhealthy air quality. Minor changes in the temperature and humidity profile could mean freezing drizzle that creates treacherous roadways instead of scattered snow showers that are a mere nuisance. Smoothed vertical profiles from a satellite may not show the true evolution of the atmosphere during a severe weather outbreak like in-situ measurements would.

Once these flights are missed, we can't get that data back. And it's not just in the United States that we are missing weather balloon flights. If we want to improve forecasting globally, we have to look globally.

Global Challenges in Weather Measurements

The atmosphere is a continuous fluid that doesn't stop based on political boundaries. Thunderstorms that originate over Africa can become hurricanes that impact the United States. Snowfall over Siberia can result in polar temperatures reaching Houston. It's important for accurate weather forecasting to have consistent data across the planet.

Unfortunately, that's not the case.

The WMO has a goal to implement a Global Basic Observing Network (GBON) - a minimum number of surface weather and upper air stations around the globe to support improved weather forecasting. Expanded weather monitoring will also support another WMO program - Early Warnings For All, an initiative designed to ensure everyone on Earth is covered by early warning systems by 2027.

There is a long way to go before there is a GBON. Using the WMO Integrated Global Observing System Data Quality Monitoring System page, we can see where weather stations are meeting GBON requirements and where additional stations and improved performance are needed to meet GBON requirements.

The above screenshots really make clear where additional surface and upper air stations, as well as improved performance from existing stations, are required. This is expensive work. Hundreds of additional upper air stations and thousands of additional weather stations would be needed in dozens of countries. To support the expansion of global weather observations, the United Nations and WMO have implemented a Systematic Observations Financing Facility (SOFF) to help finance and give technical assistance to those working to help countries reach GBON compliance. Of particular interest from SOFF is the document library, which includes gap analyses for countries, showing the status of existing weather stations and what new observing systems are necessary for countries to reach GBON compliance. The world could see 12 times more upper air data and 28 times more surface weather data if the most severe shortfalls in observations are closed.

Making the Data Usable

Collecting weather data is one thing, turning it into actionable information is another. This is a critical part of capacity development for the weather community. My team at Wisconet is working hard to turn weather data into actionable information.

Weather data needs to be used to make appropriate decisions to protect life and property and enhance the economy. Raw weather data doesn't do that. It's the tools that are developed from the data that do it. Wisconet is working on two dashboards to support decision making in Wisconsin using high resolution surface weather data - the first is for agricultural applications, and the second is for outdoor event safety.

Combining available data including Wisconet weather stations and publicly available data from NOAA, NWS, and other sources, allows us to expand our offering from just providing data maps. Adding the additional data helps growers and farmers make planting, irrigation, pest control, and harvesting decisions, and helps outdoor event planners and emergency managers make safety decisions.

Wrapping Up

Meteorologists and decision makers around the world need more data, not less. But more than that, users of weather data need tools that are specific to their use cases to help them make decisions more effectively. This requires strong National Hydrometeorological Services that have the infrastructure necessary to support global forecasting applications.