How The Weather Forecasting Works?


On a March morning at 4:55 AM, a garage door opens, breaking the morning silence in the Twin Cities forecast office of the National Weather Services. A meteorologist at the station’s Public Service Desk is launching a latex balloon filled with helium and hydrogen that is attached to an expendable radiosonde by 80 feet of thread, like they do every single morning of every single day. At the same moment, hundreds of meteorologists from around the world and the other 91 NWS radiosonde sites around the US are starting their own upper-air atmospheric soundings. These balloons will ascend up to 100,000 feet (30,000 meters) into the atmosphere over the course of the next two hours, drifting as far as 200 miles (320 kilometers) from their launch site as their payload, which weighs about the same as a grapefruit, transmits information about temperature, air pressure, and humidity back to the station.

Even if the technology is straightforward, the measurements from various altitudes that are taken every 12 hours are simply priceless. This data will be collated and integrated into unimaginably sophisticated models in the National Centers for Environmental Prediction, archived at the National Climate and Data Center, and then delivered over radio frequency and error-scrubbed by the meteorologist who originally launched the balloon. But that comes later. The Minneapolis meteorologist is currently going through their highly regimented routine when one reading catches their attention. 1,017 millibars on the barometer indicate the emergence of a high pressure front. An arctic cold front means a few more chilly commutes and balloon launches for our meteorologist. Although the data from a single balloon cannot potentially provide that conclusion, it may not signify anything to the rest of the Country or it may pose a concern. The US National Weather Service will require a tremendous amount of data to carry out their duty of informing the public about what’s to come.

The NWS first uses 160 high-resolution Doppler weather radars, each of which spends seven seconds per hour releasing brief bursts of energy and the subsequent seven seconds listening for its return to assess the presence of water or air in the sky. The fifty states, Puerto Rico, Guam, one US military base in Okinawa, Japan, and two others in South Korea all have the multi-million dollar installations, which puts the American National Weather Service in the unique position of maintaining nearly total weather radar coverage over a foreign country. There are still significant coverage gaps in the US, despite the fact that the system’s implementation was said to have increased the average tornado warning time from four to eleven minutes. The NWS lacks visibility into some portions of Tornado Alley, an area that depends on these radars to give residents there the warning they need to seek shelter when a tornado does form.

While gaps over sparsely populated western states are unfavorable obstacles to accurate forecasts, they are visible in other parts of the country. In general, these 160 NEXRAD radars do the job and serve as the backbone of the agency’s data collecting mission. But, with these NEXRAD radars now out of production and every existing one in operation, there is little room for improvement short of a pricey system-wide update. But, as revolutionary as radar technology is, it is just a substitute for measurements that are really obtained in the sky. This is why the NWS makes the effort to launch those numerous weather balloons each day, despite the fact that airplanes are already in the air. The World Meteorological Organization established the Aviation Meteorological Data Relay system because, in addition to having the sensors needed to detect atmospheric conditions, the average commercial aircraft already had these tools. 43 airlines from around the globe, including Alaska, American, Delta, FedEx, Hawaiian, Southwest, United, and UPS, have consented to gather weather data while in the air and send it to the WMO by satellite or VHF radio. Because of the high cost of using any other method to gather such frequency of atmospheric data, meteorologists regard this data as essential to the accuracy of predictions.

In reality, the loss of a large portion of this data as air traffic declined caused estimates to be worldwide noticeably less accurate at the beginning of the COVID epidemic. Nevertheless crucial high-altitude data is to forecasting what will happen on the ground, the NWS still needs to be aware of what is occurring today. There are 900 automated surface observation devices for it, again dispersed over the whole nation, mostly at airports. These include a number of sensors, such as those that measure temperature, air pressure, humidity, cloud cover, and other variables, and which automatically transmit data to the NWS as well as broadcast a radio frequency that pilots can use to learn crucial information about the situation at their destination airport. Nonetheless, there are some observations that cannot be made effectively using automated methods. Some people require human contact.

The majority of the 9,000 cooperative observer system sites around the nation are maintained by unpaid citizen volunteers who provide their time to collect data on precipitation, whether it be rain, snow, or something in between. Each adheres to a strict protocol to provide the NWS with a daily measurement of precipitation for their local area. This is the main method by which forecasters check the accuracy of their predictions; if they notice that precipitation is consistently under their forecast in a particular area, they will make adjustments, and vice versa. These data are also essential for identifying long-term patterns; for instance, it was this software that noticed that average precipitation in the US has increased 7.8% since 1901 as a result of rising temperatures making the atmosphere more capable of retaining water. As the NWS’s eyes on the ground, 350,000 trained volunteer weather spotters participate in SKYWARN, a comparable but far bigger operation. They watch severe weather so that the NWS can warn people of what is to come and so that they can validate their weather models and forecasts. These volunteers, many of whom are amateur radio enthusiasts, use their abilities to convey observations regardless of the circumstances outside. The National Weather Service needs information from sources other than US states, territories, and a few overseas radar installations since weather is a worldwide phenomenon.

The World Meteorological Organization was created for this purpose. This specialized UN organization has several duties, but one of the most important is making sure that different national forecasters can and do interact. Since a few years ago, all WMO members are required to share data, which means that the US National Weather Service (NWS) receives meteorological data from its overseas equivalents and vice versa, regardless of their geopolitical position, increasing forecast quality for everybody. The National Weather Service, one of the foremost meteorological institutions in the world, does, however, have some additional assistance in obtaining information from other countries:

satellites. A revolving crew of satellites has held the GOES-east and GOES-west geostationary locations for many years, sending a flood of pictures and meteorological data to the agency’s ground station on Wallops Island, Virginia. The GOES-16 and GOES-17 satellites, respectively, are now in charge of running GOES-east and west.

Due to a reliability issue with GOES-17, the older GOES-15 satellite is also placed next to GOES-17 to provide redundant coverage, and GOES-14 is kept powered-down in a storage orbit in case a problem with one of the other satellites necessitates a prompt replacement. If GOES east or west were to experience a coverage gap, it would affect forecast accuracy for hundreds of millions of Americans. GOES-18, which was launched in March 2022 and will shortly take the position of GOES-17 to serve as GOES-west, is in the meanwhile. The CIA also runs a non-geostationary satellite in a polar orbit to provide less often but nevertheless crucial data and pictures of the rest of the globe, whereas these are concentrated on high-frequency, continuing monitoring of the US and its near surrounds.

These GOES satellites also serve as a communication channel for the farthest-flung buoy-based NWS observation stations. Weather buoys placed hundreds or thousands of kilometers offshore offer a peek into regions of the planet devoid of people or civilization but rich in weather that may soon affect them. The network is managed by NOAA’s National Data Buoy Center, and each buoy broadcasts conditions from the Great Lakes, Gulf of Mexico, Atlantic, and Pacific in close proximity to real time. This includes providing potentially life-saving warning of oncoming tsunamis, for instance.

Yet more often than not, they supply the same information that land-based sensors do, just at a location that might help coastal communities prepare for the future. The GOES satellites transmit buoy-measured temperature and pressure to the mainland every hour, and supercomputers collect all available data to aid in the projection of an accurate weather model. This buoy, #42019 off the coast of Texas, is reading an above normal air temperature and a below average air pressure at the same time as the weather balloon hovering above Minnesota is picking up an approaching high pressure front. According to local forecasts, buoy station 42019 and other readings point to the development of a low-pressure front; hence, some rain may soon be expected around the gulf.

A little rain may soon be the least of the NWS’s concerns, though, as weather models predict, taking into account both an arctic front descending through the midwest and a warm, moisture-laden front emerging across the gulf. At many various sizes and with very distinct outcomes, weather evolves.

The NWS must forecast meteorological events on a variety of scales in order to safeguard the individual, the thousands, and the millions. The National Centers for Environmental Prediction’s Weather Prediction Center, where information from balloons, buoys, aircraft, and satellites all around the world is fed into various weather models, is where most of these forecasts of every size and shape originate. Although though the NWS creates a wide range of unique models, they all essentially function the same. In each, data from observed environmental circumstances are entered into a variety of equations; the solutions, when shown on a map, provide an approximation of the current atmospheric conditions as well as a forecast of how weather patterns will evolve and move in the future.

All weather modeling is based on the same data, logic, and physical principles; nevertheless, what sets one model apart from another is its resolution and range. With its 2 mile (3 km) resolution, real-time visualizations, and hourly updating, the High Resolution Rapid Refresh model, for instance, offers unmatched detail and accuracy on the development of rapidly moving thunderstorms over the coming hour, giving meteorologists more confidence in issuing severe weather warnings. Its breadth is too narrow for it to be able to zoom out physically or temporally to foresee national changes even hours in the future.

It can advise you to go right away, but it can’t advise you whether or not to pack a raincoat for tomorrow. Meteorologists utilize the North American Mesoscale model, which has a lesser resolution than the HRRR but can identify trends and anticipate their courses up to 61 hours, to assist in making such crucial decisions as whether to wear sunscreen or rain coats.

The Global Forecast System, which again adopts a wider perspective than the NAW but is able to interpret atmospheric patterns and weather developments on a global scale and to predict out to 10 days in the future, is used by the NWS to get an even bigger picture. This is important in tracking the lengthy development of hurricanes, for example. These models work as extremely potent forecasting instruments when combined. Moreover, extremely powerful computers are needed to run all of these simulations. Two of the US’s biggest supercomputers are what enable all of these models. Congress approved more funds for NOAA because the NWS’s Global Forecast System failed to predict Hurricane Sandy’s potential for harm. As a result, NOAA said in 2016 that the installation of two supercomputers in Reston, Virginia, and Orlando, Florida, had improved its computational capability by a factor of ten.

The $45 million investment gave NOAA access to the 18th fastest computer in the US, leveling the playing field between the NWS and its European rival. In 2018, NOAA doubled down on the modernization initiative by investing in its supercomputers once again, increasing their storage by 60% and their computational power by more than tripling. These investments allowed way for upgrades and enhancements to its models, from the HRRR to the GFS. Importantly, and further justifying such investment, these supercomputers are also tasked with creating ensemble models, which essentially involves running the same models repeatedly with the parameters slightly changed to reflect the random chance of real-world weather and to give forecasters an idea of the level of uncertainty for a given prediction. The forecast is more definite if it holds true despite the random chance that the ensemble model has thrown in, and vice versa. All of these models are incredibly sophisticated and very costly, as are the supercomputers that run them continuously. They have already shown to be of great use when dealing with extreme weather, and they will do so once more when distant, dissimilar signals of storms silently form. On an early March day, hours after a weather balloon was launched in Minnesota and a buoy was uploaded off the coast of Texas, they were proving their value once more as GFS forecasts and then ensemble GFS forecasts started to project with a higher degree of confidence that a storm with significant destructive potential might be developing.

The GFS models had already detected the early beginnings of what meteorologists refer to as a Gulf Low when millibars plummeted across Texas. They can now forecast its course and possible intensity thanks to buoys that have proven such forecasts to be accurate. At the same time, a lowering cold front has also been developing. The deepening cold front and a U-shaped gulf stream mean that this particular storm doesn’t appear to be moving along the western edge of the Appalachians, but rather through some of the most densely populated areas of the American east coast. This Gulf Low appears to be mirroring historical patterns in that it too is poised to move a mountain of moisture out of the Gulf of Mexico. Snow, and a lot of it, is moving toward the northeast, according to GFS forecasts.

In obviously, predictions alone alone have no value. The National Weather Service must genuinely explain what’s to come in order to prevent the catastrophe of an unprepared people in the path of an imminent winter storm—plows still in the parking lot, businesses without additional stock, children still in school, commuters on the road at the wrong moment.

To do this, the organization has 122 weather prediction offices spread out across the fifty states, Puerto Rico, and Guam, each in charge of a certain geographic region of the nation. A dozen employees at each location work round-the-clock to evaluate models and provide forecasts for the area they serve. Probably more crucially, though, local WFOs are responsible for monitoring and sending out alerts for the most immediate and frequently most hazardous dangers. These include severe thunderstorm, flash flood, and tornado warnings, among other things. WFOs may also customize their operations to meet the needs of the populations they serve thanks to their on-the-ground presence. The adjacent Denali—tallest America’s mountain—is the subject of a daily climbing weather prediction from the Fairbanks, Alaska WFO. Hundreds of kilometers of wheat fields surround the Dodge City, Kansas, WFO, which provides daily soil temperature information to farmers so they may maximize harvests.

The Grand Junction, Colorado WFO keeps track of and reports on fire weather conditions, notifying the wildfire-prone region it is in charge of to the daily danger of one beginning and spreading and resulting in the notification of campfire bans, power grid shutdowns, and other things. The Grand Junction office, like the majority of WFOs in fire-prone areas, has an Incident Meteorologist on staff who is ready to deploy quickly to the field and embed in a wildfire camp to issue extremely precise forecasts used by aerial firefighting pilots, hot shot crews, and other firefighters to approach and attack the burn safely.

Last but not least, many of the NWS’ ground activities are carried out by the regional WFOs, who also manage the regional NWS Twitter account, maintain doppler radars, automated surface observation stations, and much more. Yet in the end, certain weather events are simply too large or extraordinary for tiny, local WFOs to handle. There are national offices for it. Some are simple to understand. There is an additional system of thirteen river forecast centers that roughly divide the lower 48 states and Alaska according to watershed and issue forecasts for their rivers’ flows and, most importantly, their flood risk. Rivers, for example, run for hundreds or thousands of miles through countless WFO zones.

The National Hurricane Center, located in a bunker-like structure in Miami, Florida, is perhaps the most well-known of these national offices. It is built to withstand and continue to function during a category five storm, which is the strongest possible. Predicting the routes and intensity of storms is a difficult problem for them. Lives will be saved if they get it right and people are ready. Lives will be lost if they don’t follow the forecast, and they might be doing all of this while a storm passes straight overhead. Due to the way the system is set up, if the Miami center GOES offline during a storm, the National Weather Prediction Center in College Park, Maryland is ready to take over immediately.

There is also a second office in Honolulu, Hawaii, which staffs up if a relatively uncommon central Pacific hurricane forms. The National Weather Service has the unusual responsibility of predicting the weather not just outside the borders of the United States but even outside of our planet. To report on and anticipate future solar radiation and other space weather, the NWS’ Space Weather Prediction Center in Boulder, Colorado uses data from GOES satellites and terrestrial sensors.

High solar radiation activity, for instance, can have an influence on the human world, so this isn’t just for fun. Commercial airplanes cannot fly routes over the poles where they rely on high frequency radio to communicate with air traffic control if the center has anticipated excessive levels of radiation. And that’s only the beginning of the problems solar flares may bring.

The National Weather Service’s work ultimately boils down to a couple megabyte XML data feed updated every few minutes, despite the thousands of hourly observations it receives from around the globe and beyond, the full processing power of two incredible supercomputers, and the combined efforts of thousands of highly-trained personnel.

Often, individuals don’t acquire their weather information directly from the National Weather Service; instead, they use their phone’s weather app, a TV meteorologist, or another method that suits their needs. Even if it is repackaged and marketed, nearly every source of meteorological information in the US depends on the work of the National Weather Service, therefore the NWS strives to make this data widely accessible to those who deliver it to the end-user.

Everybody, from individuals to commercial companies, is legally allowed to use these data since, as a product of the US-government, it is not copyrighted and does not require a license. These XML data-feeds are a major source, providing a structured set of current conditions and forecasts that can be quickly adapted and interpreted by software.

Of all, when communication is most difficult, that’s also when weather information is most crucial—the internet and TV could not be available when it counts. Hence, a NOAA Weather Radio station, which automatically broadcasts a 24/7 feed of conditions and predictions, is within reach of around 95% of the US population.

The Emergency Managers Weather Information Network, which is managed by the National Weather Service as a last option, employs its GOES satellites to broadcast a data-feed down to earth that is, uniquely, accessible to anybody with a suitable satellite receiver. TV stations, emergency managers, and others can access this crucial information regardless of the state of the environment thanks to this data-feed, which is accessible at 1694.1 megahertz.

It includes all the highlights of the conditions and forecasts for locations across the US, as well as these visual forecasts, satellite imagery, and other images for onward use. In summary, the National Weather Service uses a few satellites and supercomputers, hundreds of radars, observation stations, buoys, weather balloons, aircraft, and weather forecast offices, as well as thousands upon thousands of committed employees and volunteers to warn the east coast of impending snowstorms.