With thousands of flights grounded, weather forecasters were sounding alarm bells about lost data that could harm forecast accuracy. Sensors aboard commercial aircraft gather data that augments information from the weather balloon network at observation sites around the world.
Depriving computer models of the aircraft data threatened to erode forecast accuracy for all the major models in use, from the vaunted European model to the main U.S. model, known as the Global Forecast System, or GFS.
Spire, which has 88 tiny satellites, each the size of a loaf of bread, in low Earth orbit, gathers what’s known as radio occultation data, which can be used to develop profiles of moisture and other properties of the atmosphere. To help plug any aircraft-related holes in data gathering, the company offered its data free to the European Center for Medium-Range Weather Forecasts and the United Kingdom Met Office.
So far, it would seem those agencies are pleased by the proof of concept, although it’s not clear the data has been a game-changer for them. In a statement, John Eyre, a fellow at the Met Office, said radio occultation data helps improve weather forecast accuracy in general.
“We have used RO data from other satellites for many years, and we know their value for improving our weather forecasts. The offer of data from Spire is very welcome. It will make a valuable contribution to mitigating the loss of other weather observations during the COVID-19 period,” Eyre said, referencing the disease caused by the coronavirus.
Spire’s satellites, which it has named LEMURs, for low-Earth multiuse receivers, also collect maritime and aircraft tracking information, providing the company with another source of revenue.
Radio occultation can be thought of as akin to a space-based scan of the atmosphere, and it allows meteorologists to gain observations of atmospheric temperature and moisture using a satellite receiver that measures signals sent by GPS and other navigation satellite systems.
The signal between the two satellites gets refracted, and information about the temperature and water vapor content of the atmosphere can be gleaned from the magnitude of this refraction.
NASA deal marks a key step for Spire, and for similar companies
Now the company, which has raised $150 million in five funding rounds, has taken another step that puts it ahead of some competitors in the race to supplement weather data gathered by far larger and more expensive government-procured satellites. The company inked a one-year, $7 million deal with NASA to provide it with radio occultation data, which will be available to agency scientists for use in their Earth-observing missions.
The contract is part of a NASA pilot program to evaluate how Earth science data from commercial small-satellite systems could supplement observations from the agency’s satellite fleet. Spire, which has offices in six locations (including Luxembourg; Boulder, Colo.; and Singapore) was awarded a contract for test data in September 2018, along with Planet and Maxar.
Platzer said the combination of deals with the U.K.’s weather agency, the European center and now NASA signal “a growing-up moment for the small-satellite industry contributing to weather prediction in a very meaningful way.” Platzer said in an interview that just a few years ago, his company’s tiny satellites were dismissed as “a toy.”
The test data, and now the contract through NASA’s Commercial Smallsat Data Acquisition Program, is ultimately aimed at improving the information available to the agency and potentially forming a different model where, instead of building large, multibillion-dollar satellites, the agency contracts with companies to use fleets of lower-cost satellites that can provide similar or superior information.
A NOAA decision looms
Like NASA, the National Oceanic and Atmospheric Administration has evaluated Spire’s radio occultation data using an initial private-sector project. Platzer said that the agency’s feedback was that it would like to see more data volume from Spire’s network and that the company has since focused on generating more information from each small satellite it launches.
On Friday, NOAA released its evaluation of the initial private-sector project, known as the commercial weather data pilot. The agency concluded that, based on trial data from Spire and its main competitor, GeoOptics, the commercial weather sector “is able to provide radio occultation data that can support NOAA’s operational products and services.” Data from these companies showed a potential to improve weather forecast accuracy, particularly for near-term forecasts.
Before the end of the current fiscal year, NOAA is scheduled to announce a draft statement of work and request for public comment for its first acquisition of commercial space-based radio occultation data for operational use. The NOAA budget contains $5 million for this data purchase in fiscal 2020, with the president’s budget requesting $15 million in 2021.
This is a tiny fraction of the $1 billion-plus that NOAA spends on its weather satellites each year, but it represents the agency dipping its toes into the rapidly growing private-sector space industry.
Spire offered NOAA the ability to tap into its occultation data free during the coronavirus-related air travel slowdown, but the agency was unable to move as quickly as its European counterparts, Platzer said. He said this may be because of the structural difference between NOAA and the European center; the latter is singularly focused on producing medium-range forecasts, whereas NOAA manages fish stocks and operates weather satellites and the sprawling National Weather Service, among other functions.
“That is a structural difference, a structural burden that the NOAA administrator has,” Platzer said.
One dilemma that NOAA faces with moving forward on purchasing weather data from a private company such as Spire is that the agency will still be on the hook for providing accurate weather forecasts regardless of what happens with the contract. So if Spire’s satellites were to fail or the company’s data turned out to be unreliable, then NOAA would still be responsible for finding a way to avoid losing accuracy.
“Ultimately, NOAA must have confidence in the commercial sector to be able to use their capabilities to serve our mission, which we must meet with or without commercial data contributions,” NOAA spokesman John Leslie said in a statement. “NOAA looks forward to working closely with the private sector to meet its mission.”
Shali Mohleji, a NOAA policy adviser in the Obama administration, said contracting with private firms comes with pros and cons for NOAA.
“I think that the responsibility to provide regular and reliable forecasts remains with NOAA, so the risk of losing data remains a risk to the agency,” she said in an email. “The new arrangement doesn’t actually ‘offload’ this risk, it just converts it to a critical supply chain risk, and one that is likely to be harder for NOAA to monitor and mitigate.”
Spire’s main rival for the upcoming NOAA data contract is GeoOptics, which launched its first satellite in January 2018. It is headed by former NOAA administrator Conrad C. Lautenbacher and was also awarded a contract under NOAA’s pilot program to evaluate the accuracy and uses of private-sector occultation information for weather forecasting.
In April, GeoOptics announced the launch of a processing system for its radio occultation data, which it says helps improve its accuracy and the speed at which it reaches users.
Spire and GeoOptics aren’t the only providers of occultation data. A publicly funded program between the United States and Taiwan known as Cosmic-2 gathers such data as well. Spire and GeoOptics maintain that even more occultation profiles would be beneficial for forecasting and other uses.
Platzer said Cosmic-2 is limited to the equatorial regions and generates 5,000 to 8,000 radio occultation profiles a day, whereas Spire’s network can generate more than 10,000 profiles a day and cover the whole globe.
Questions about data quality and value
Outside experts unaffiliated with Spire said that while it is important that agencies such as NASA and NOAA engage with the rapidly growing private-sector satellite industry, it’s not clear how big the upsides will be. “We don’t quite know what the additional value of Spire or GeoOptics [data] will be yet,” said Frederick Carr, a professor emeritus at the University of Oklahoma.
Mary M. Glackin, who serves as president of the American Meteorological Society but spoke in her capacity as a former executive at IBM’s the Weather Company, said Spire is noteworthy because it is also developing its own weather prediction model, which would allow it to get a clear internal idea of how its satellites are performing.
She said that with any private weather company, it is important to question the value of the data. Companies tend to think their data is worth more than it really is, she said. For example, if an increase in the number of radio occultation profiles allows the Weather Company, which operates the popular website weather.com, to improve its three-day forecast by 2 percent, “how much more money can you really make on that?” she said. “It’s not at this point enabling whole new business lines or anything like that.”