Why the National Weather Service Should Not Be Privatized


Over the past thirty years, the overarching debate about the proper role of government has come to include a governmental agency whose function had not been questioned previously: the National Weather Service (NWS). Since 1983, the idea of cutting taxpayer funding for the NWS and related agencies, and turning over their operations to private companies, has periodically surfaced in Congress and the media. The proposal has taken two primary forms: cutting funding for forecasting operations with the expectation that private firms would take over the task, and selling satellites or other sources of weather data to the highest bidder and buying back the data.

The NWS began in 1871 as a military agency and has operated in three different government departments after becoming a civilian agency. Still, one thing remained constant: Its operations were uncontroversially agreed upon to be a proper function of the government. While conservatives and liberals argued heatedly about the role of government in New Deal operations in the 1930s, a series of killer hurricanes—including a 1935 storm that devastated the Florida Keys and killed as many as 600 people—directed debates about the agency only in the direction of how it could prevent such tragedies from happening again. The focus was on improvement of the agency[2].

However, toward the end of the twentieth century, a new idea appeared. John Byrne, the director of the National Oceanic and Atmospheric Administration (NOAA) under President Reagan, proposed in 1983 to cut costs by selling US-owned weather satellites to the highest private-sector bidder, buying back the data the satellites provided, and dismantling the forecasting arm of the National Weather Service. The private sector, Byrne said, could provide forecasts to the public—and would do a better job at it than the federal government. However, the public did not agree, and the highly unpopular proposal was tabled soon after it was floated[7][14].

Throughout the 1990s, the notion of privatizing all or part of the National Weather Service (or related agencies) stayed on the outskirts of political discourse despite a Congress with strong philosophical sympathies. Privatization talk was kept down by a vibrant economy, a popular president opposed to privatizing the agency, and a decade unmarred by any exceptionally large loss of American lives in a weather event, including the well-forecast Category 5 Hurricane Andrew. During the first half of the 2000s, a strong focus on war and national security and a continued reprieve from highly deadly weather events also kept privatization from coming up in discourse.

However, in 2005, after the devastation wrought by Hurricane Katrina, the idea of privatizing the National Weather Service became a topic of public debate once more. Senator Rick Santorum (R-PA) introduced a bill (S. B. 786, 109th Congress, 1st Session) that would have prohibited the agency from publishing forecasts directly to the public except for urgent severe weather events. The agency would have had to provide its "routine" forecasts to private-sector companies instead. The bill placed no restrictions on whether the private firms could then charge the public to receive the forecasts that their tax money had paid for, nor did it institute requirements for the private purveyors even to be trained in meteorology[12]. Therefore, in addition to putting the public in a position of potentially having to pay for weather forecasts, the bill could have put hundreds of highly credentialed scientists out of a job with no guarantee that they could gain employment in the private sector.

Santorum harshly criticized NOAA, accusing the National Hurricane Center of doing a poor job of forecasting Katrina. Other Senators from his own party and public officials from the affected states of Louisiana and Mississippi disputed that claim[5]. The meteorology community largely rallied against the proposal and in support of the NWS[13], uncovering evidence that Santorum had received $50,000 in campaign contributions from executives in a large privately owned Pennsylvania weather firm and intimating that this was his motivation for introducing such a bill[1]. Santorum lost his Senate seat in the 2006 election and the control of Congress flipped.

The idea of privatizing-by-proxy arose once more in 2011, when the House of Representatives proposed a budget that would cut funding for the National Weather Service by 30 percent. The budget also included cuts for the National Hurricane Center, the aerial reconnaissance operations into hurricanes, and the National Buoy Data Center[6]. Supporters of the proposal, including the private firm AccuWeather (which had donated the $50,000 to Santorum), defended the cuts by saying that the private sector outperformed the NWS. A controversial pro-privatization editorial by bloggers Iain Murray and David Bier was published on the website of Fox News[8], and after criticism from NWS supporters, they defended their piece in the climate-change skeptic website "Watts Up With That"[9]. Their writing has become the primary source of arguments for adherents of privatizing the NWS and related agencies.

Forecast Accuracy

Arguments in favor of privatizing weather operations usually originate from opposition to what they claim is "big government." However, adherents also advance arguments based upon testable claims. These claims revolve around a cost-benefit analysis involving the efficiency and accuracy of the agency's products versus what is spent to generate them.

The federal government spends almost $1 billion for National Weather Service operations alone[4], a figure that does not include maintenance of many satellites or oceanic observations. The agency is the single biggest expenditure in NOAA. Privatizers argue that forecasting the weather is an unnecessary government function, because the private sector could do the same job more accurately. They argue that a free market encourages competitors to improve their products, a basic application of capitalist theory to the product of a weather forecast.

In fairness, some historical support does exist for this claim. In the 1950s, the Weather Bureau did not issue any sort of warning for tornadoes. Forecasting was not at a level of expertise in the Bureau at which such warnings could be issued with confidence, and the Bureau feared a "cry wolf" scenario of false alarms. However, the Tinker Air Force Base in Oklahoma was issuing tornado warnings for itself. Television meteorologist Harry Volkman received the base's information and issued on-air tornado warnings for civilian areas as well. Officials in the NWS were not pleased, but Volkman's employer and the Air Force base supported him. The Weather Bureau finally had to relent and begin issuing warnings for tornadoes, which meant that their forecasting had to improve[7].

However, it is important to point out that privatizers would exclude one "competitor"—the NWS—from the pool by dismantling it, which was not the case in the 1950s. It should also be noted that nothing whatever prevents an existing forecasting company from competing with the NWS. There is no prohibition on a company from developing its own forecasts, and many do. And finally, competition does exist in the NWS itself—between different local offices—and research has shown that such competitiveness leads to forecasting and warning improvement[11].

Proponents of privatization make a specific claim about competition that is, in theory, testable: the claim that the private sector already produces better weather forecasts than the National Weather Service. As it turns out, however, this claim cannot be shown to be true. When Santorum introduced his 2005 bill and made his criticisms of the NWS and National Hurricane Center, members of the meteorology community largely came out strongly against the idea. Santorum's claims that the Hurricane Center did a poor job of forecasting Hurricane Katrina compared to the large Pennsylvania firm AccuWeather were immediately tested and found to be patently untrue[13].

The claim that the private sector generates better general forecasts than the National Weather Service is not borne out either. In the first place, determining forecast accuracy is inherently subjective, since it depends upon the importance given to a variety of factors. Most forecasts are considered good if the temperatures for a given day are close to those predicted. However, on days when precipitation occurs, that factor is given more significance because it is more disruptive. Severe weather events are given greater importance still in personal assessments of accuracy; a forecast that correctly predicted temperature and rainfall but completely missed a tornado outbreak would definitely not be considered good. The fuzzy nature of forecast accuracy has enabled virtually any claim to be made by selectively weighting data. However, a 2007 study that looked at the most basic of variables—the daily temperature—found that the source with the best forecasts varied. Private firms generally produced better same-day temperature predictions than the NWS, but the NWS did better one to three days in advance. Farther out than that, private firms beat the NWS again[3]—though largely because the NWS did not issue forecasts beyond a seven-day period at that time. However, privatization proponents' claims that the private sector unambiguously beats the NWS were not supported in this aspect of forecasting, which, again, did not even examine more disruptive weather. It is worth noting at this point that most private firms do not even issue their own time-critical severe weather warnings, but simply repeat those of the NWS. Potential legal liability (or bad press) for injury, death, or property loss after a missed warning has perhaps been the deciding factor for private firms against issuing these warnings.

Impact on Research and Enterprise

Moreover, scientists and small-scale entrepreneurs in the weather community have an additional set of concerns regarding the consequences of privatization. Most NOAA weather data are available free of charge with few restrictions on usage. Scientists use the data in their research. Private firms also use the data, perfectly legally, to make a profit. The broadcast meteorology sector uses NWS data for its own products, which make money from advertising revenue. Smaller private firms such as Base Velocity LLC (developer of the popular paid mobile app RadarScope) and Gibson Ridge Software (developer of the scientific radar software GRLevelX) take weather model and observation data and repackage them to be useful and visually attractive to weather watchers, but the data that the programs read are from NOAA.

However, if a private company were to gather weather observations, those data would become the intellectual property of that company. Indeed, the products that private firms create from repackaged or modified NOAA data are considered protected intellectual property, even though the original sources are not. The possibility of fees to use the data in research would arise at once, and with that would come attendant concerns of additional funding requirements for any such research—in an era in which funding is difficult to come by. The existing private sector would be negatively impacted too. Although John Byrne's 1983 proposal would have had the government purchase satellite data back from the firms that maintained the satellites, this type of contract would not have to stipulate that the government could then redistribute the data for free with no restrictions attached. Indeed, the private firm in the contract would not want this clause, as it would allow competitors to use that same dataset without paying for it. Any entity, corporate or individual, seeking to generate a new product would likely have to either buy the data or acquire its own, a change from the current model that has led to such a vibrant private sector of weather products.

The acquisition of data is itself an argument against privatization. The general public already receives much of its weather information from privately owned sources, privatizers argue, primarily mobile apps and TV weather people. With the advent of a nationwide network of personal weather stations, public participation in data-gathering, broadcasting firms, and weather websites, much of the process has already been privatized, they say. However, generating a good weather forecast requires a vast network of sources far greater than the personal weather stations and mobile app reports that adherents call "privatization." In fact, the use of personal data is not a new phenomenon at all. Individuals made such observations at home in the 1800s, but it was not nearly enough data to produce an accurate forecast, nor were the observations necessarily reliable. The modern network includes highly calibrated ground stations, radar domes, satellites, aircraft reports, buoys, and weather balloons to acquire accurate observations. (Atmospheric scientists today are intrigued by the possibility of incorporating spot reports volunteered by mobile users into weather modeling, but under no circumstances are scientists considering replacing high-quality full-atmospheric observations with surface-only reports of unknown accuracy.) Furthermore, a forecast requires more than just data-gathering. NOAA also funds million-dollar supercomputers to run the numerical weather models and teams of experts who can focus on small areas when urgent forecasts or warnings are needed. The sheer scale and expense of the infrastructure requirements locks out most organizations from ownership except the federal government and large companies—which would then have a private monopoly.

The problem of data ownership in a privatized National Weather Service is global. While some data sources, such as satellites, can acquire observations from around the world, other types of global observations cannot be gathered by NOAA because of national sovereignty. However, since the atmosphere is global and the principles of fluid dynamics dictate that changes in one part will affect a part elsewhere, it is nonetheless crucial for global data to be available for forecasting. The World Meteorological Organization (WMO) is a consortium of governmental agencies involved with atmospheric science, and through the WMO, data-sharing occurs across national boundaries. The WMO certifies data sources—including those maintained by NOAA—as being "official" for scientific purposes. While it could certify data from a private source, the seal of WMO certification itself would provide a distinct advantage to any firm that obtained it, further encouraging a monopoly to develop.


In conclusion, supporters of privatizing the NWS and related agencies have made two hard claims, that privatization would save money and that the private sector produces better forecasts already. The latter assertion, however, cannot be demonstrated to be true (and for some events has been shown to be demonstrably false), and the invocation of the more than $1 billion spent on all weather-related operations is balanced against a consideration of what that money buys, which is much more than just weather forecasts, but a pool of data that supports research and an existing free market that any entrepreneur could enter at a very low cost. Perhaps because they ultimately originate from an objection to the government's participation in any activity that could be commercialized, arguments for privatization have been focused primarily on the supposed benefit to the "consumer," meaning the average person who checks weather forecasts, and have virtually ignored the extremely significant impact on those in the scientific and enterprise community who are not merely consumers of the data, but also producers of new products from it.

Essentially, the data generated by NOAA have become a public resource, but with the benefit of being virtual and therefore not exhaustible. Scientific research takes place and a healthy private sector already exists because the majority of weather observations produced by NOAA are in the public domain. Furthermore, no law exists that would prohibit a company from gathering its own private data, nor from running its own supercomputers to generate numerical forecasts. Turning a public resource into a privately owned store of intellectual property would, ironically, be potentially very detrimental to a free market built upon competition, and would cast a chill upon future atmospheric science research. For these reasons, the operations of the National Weather Service and its related agencies should not be defunded and privatized.


  1. Citizens for Responsibility and Ethics in Washington, 2006: "Beyond DeLay."
  2. Davies, Pete, 2000: Inside the Hurricane: Face to Face with Nature's Deadliest Storms. Holt, pp. 71-73.
  3. Hansen, Brandon U, 2007. "Internet Weather Forecast Accuracy."
  4. Ingram, Frederick, 2013: "National Weather Service Information."
  5. King, Bob, 2012: "Rick Santorum's campaign could be clouded by 7-year-old attack on National Weather Service."
  6. Klaustermeier, Kayleigh, 2011: "National Weather Service budget cuts." MyArkLaMiss.com, Feb. 25, 2011.
  7. Mathis, Nancy, 2007: Storm Warning: The Story of a Killer Tornado. Touchstone, pp. 84-86, 90-93.
  8. Murray, Iain and David Bier, 2011: "Do We Really Need a National Weather Service?"
  9. Murray, Iain and David Bier, 2011: "Government Funding of the National Weather Service: A Response to Our Critics."
  10. National Oceanic and Atmospheric Administration, 2007: "The History of Numerical Weather Prediction."
  11. National Weather Service, Birmingham, AL, 2014: "National Weather Service Birmingham AL Reduced Tornado Warning False Alarm Rate By 31% Since April 2011."
  12. National Weather Service Duties Act of 2005, S. 786, 109th Cong., 1st Session (2005).
  13. Thunstrom, Tom, 2005: "The War Between Accu-Weather and the NWS Flares Again."
  14. Wagman, Robert, 1986: "Privatization is a major issue in Reagan's second administration." Ocala Star-Banner, Jan. 6, 1986.