Weather is described as the current state of the atmosphere. It is always around us in the form of clouds, rain, wind, and air temperature. Yet, predicting the weather beyond 7 days into the future is excruciatingly challenging. We as a society still have issues not being able to forecast correctly. But why? How have we as a species come this far with sending humans to the moon and prolonging life via medicine, yet still are so wrong about what the conditions will be in the next 24 hours? It all comes down to what makes a forecast in the first place.
Observations
All around the world, the weather is continuously being monitored. Weather is recorded with human observations, automatic systems on the ground, and satellites in space that constantly take images of the Earth. These observations serve as the foundation for the forecasting process. These observations of temperature, wind speed/direction, air pressure, moisture content, etc. are then turned into numbers, called Numerical Weather Predictors (NWP). An NWP is a value that relates to current weather, such as 1004hPa or 25knot winds. These NWPs are then collected by both private and public weather forecasting and plugged into special equations.
For the ocean, most of these observations are recorded by voluntary observing ships and buoys. The Voluntary Observing ships have been steadily decreasing in numbers since the 1970s however because of the costs of operation for the merchant companies. Buoys on the other hand have become more advanced. There are two main types of buoys, moored and drifting. As the names suggest, drifting buoys move along with the ocean current, while moored buoys stay stationary. These observation methods allow scientists to observe weather over 71% of the world’s surface that is covered by the oceans.
Models
Once the NWPs have been collected, they are inserted into climate and weather models. These models use weather observations in the form of NWPs and plug them into advanced mathematical equations to predict the future state of the atmosphere. The different use of spacial reasoning, atmospheric dynamics, and output resolutions are what give the models their variance. There are, of course, limitations to these models.
Weather Forecast Model Websites
https://www.tropicaltidbits.com/analysis/models/ – Forecast Models
https://www.pivotalweather.com/model.php – Forecast Models
First, if there were to be a perfect model, it would need the initial state (temperature, velocity, pressure, etc.) of every single molecule of air all across the world. Well, a single breath of air contains around 10 sextillions (10^21) molecules. Now picture how many molecules are over the entire atmosphere. The number of molecules is functionally infinite, and the act of measuring all of them is literally impossible. Instead, weather models use averages over large areas in order to make the numbers quantifiable. Secondly, suppose humanity had the ability to gain the initial values of every single molecule of air in the world. Well, computing power would be needed to calculate the equations based on these initial states. The National Oceanic and Atmospheric Agency (NOAA) actually owns some of the fastest computers in the world for the very purpose of utilizing weather models.
The Future
The limitations listed above give the reasons why there are always going to be mistakes in the weather prediction done by computers. Forecast meteorologists compare the knowledge they have acquired from studying weather patterns to the outputs from the weather models. New technologies such as Artificial Intelligence and Quantum Computing decrease the errors of these models, yet at the end of the day, it all comes down to the knowledge of a human to figure out the true picture of the weather. Real-world experience and an understanding of atmospheric dynamics is still the best predictor of future temperatures, winds, and waves. So please give your weather forecasters some slack. At the end of the day, they are trying to perform a truly impossible, chaos-filled task.