Weather Prediction - Behind the scenes

Are you planning your holidays or simply planning for a day? What is one of the most important planning we do these days before packing our bags or getting ready to step out of our home? The answer is simple "Weather Report", to find out further answers to - Do I need to carry an umbrella or apply sunscreen or is it a good day to ride a bike and many more. However, have we ever thought about how these weather reports are prepared? It is not a matter of years or decades but centuries where humans have applied different techniques and methods to simplify the tracking of the most complex phenomenon of our mother earth. The evolution of technology and usage of science-based weather forecasting techniques has made us rely on and get alerted on how to plan our days ahead of time.?

In the 19th century, local weather data was used for science-based weather forecasting but today with emerging technologies, fluid dynamics and thermodynamics are used to measure and forecast. To make a better and more accurate prediction, the meteorological department had deployed thousands of devices and instruments all over the world including space which are constantly recording and transmitting data in huge amounts and to process them we need heavy computational machines. Therefore, many departments use supercomputers which could be a million times faster than our average home desktop to process such a large amount of data. And ground truth for the success of any weather forecasting system depends on the data it has collected to make accurate predictions well in advance.

It is not only individuals who need to plan but many critical businesses are relying on these weather predictions such as:

• Agriculture - Unsuitable climatic conditions can make a high impact on crops.?
• Flights - You might have experienced a flight delay because of bad weather conditions.
• Expeditions - Every mountain expedition-hosted country would like to make sure that reliable and safe weather information must be provided to their mount climbers.

Data Collection:

To make a weather forecast, an organization needs to collect millions or billions of data points to maximize its chances of making a significantly accurate prediction. It is done by collecting the current state of our atmosphere such as relative humidity, temperature, wind speed, moisture etc. which can help to predict how our atmosphere can possibly evolve for upcoming hours or days.

Today most of the data is collected for weather prediction using IoT sensors across the globe and the meteorologist department had been using various equipment to collect meaningful data with advanced technology so that meteorologists can make better and faster predictions than ever before. Let's take a look at some of the most common methods to collect weather data points:

Weather Instruments

Following are the different weather instruments which are used to measure weather conditions:

• A barometer measures air pressure
• A thermometer measures temp
• An anemometer measures wind
• A rain gauge measures the amount of rainfall
• A hygrometer measures humidity
• A snow gauge measures the amount of snow

These weather instruments are installed at weather stations to help collect the weather data. Many of these instruments are placed on land and also on the ocean. However, to collect weather data for the atmosphere, the below-given weather devices are used:?

• Doppler radar - These radars can detect all types of precipitation, rotation of thunderstorm clouds, tornado and wind strength for observing severe storms.

• Satellite Data - Weather satellites monitor Earth's weather from space and collect observational data points for meteorologists to analyze the weather. There are mainly 3 types of weather satellites:

1. Polar-orbiting - satellites orbit close to the Earth's surface and take multiple images per day
2. Geostationary - satellites stay over the same location and take frequent images of the Earth's surface (details to minutes or seconds)
3. Deep Space - satellites face the sun to monitor powerful solar storms that can also impact Earth's weather

These satellites help transmit aerial views of clouds, volcanic activities, dust storms, pollution and many more. It also helps capture the non-electromagnetic waves such as infrared which is further used to show the temperature graphs of our earth.

• Radiosondes - You must have seen in many movies and documentaries where meteorologists release balloons that carry a device called Radiosondes. These devices are the primary source of upper-air data which floats to the upper stratosphere where it collects and sends back data every second about air pressure, temperature, relative humidity, wind speed and wind direction. During severe weather, frequent balloons are launched to collect additional data about the weather. On average 900 radiosondes are released twice a day across the globe.?

• Automated Surface Observing Systems (ASOS) - These systems are installed on Earth's surface to monitor the sky condition, surface visibility, precipitation, temperature and wind speed. These systems also provide data related to the amount of rainfall and snow in a certain location.

• Buoys - These devices are placed on oceans to do the atmospheric measurement. Thousands of weather buoys can be found drifting in the world's ocean. The instruments installed on Buoys can measure both above and below ocean surfaces. And I am sure many of you have heard about El Ni?o and La Ni?a ocean trends and ocean buoys are quite helpful in capturing enough data to predict these oscillation trends.

• Aeroplanes - Did you ever realize that aeroplanes (passenger or cargo) are used to capture atmospheric data points? Today, aeroplanes are equipped with weather devices to gather and transmit weather data that are then used for improving weather forecasts issued worldwide. And do you know where these devices are placed? Have you noticed the pointy nose of an aeroplane and must be thinking of leg room for pilots LOL. Well, the devices are placed in the pointy nose. Next time, if you see the aeroplane nose then you know how useful it is for weather prediction.

How do these data points get converted to Prediction?

Metrological departments use weather models to convert these data points into meaningful and to make significantly accurate weather predictions. These weather models are generally referred to as numerical weather predictions which are the core of today's weather forecast system. These weather models power all the forecast information you see.?

These weather models are a simulation of the future state of our atmosphere. Millions of observations are used to perform billions of calculations to produce a 3-dimensional picture of what the atmosphere would look it in the future. And to do these calculations massive computational-powered supercomputers are used to stimulate the entire global weather forecast.

There are 2 types of weather models:

Global models - produce a forecast for a week or two for the entire globe. Generally, these models have a low resolution because of wide coverage ànd fewer data points for a given area. Below given are some of the most powerful and well-known Global models:

• ECMWF - European Center Medium-Range Weather Forecast (Resolution ~9 to 14 KM)
• ICON - Icosahedral Nonhydrostatic (ICON7 and ICON13) (Resolution ~13 KM)
• GFS - Global Forecast System (Resolution ~23 KM)
• GEM - Global Environmental Multiscale Model (Resolution ~22 KM)
• UKMO - United Kingdom Multimodel (Resolution ~10 KM)

Regional models - produce a forecast for a few days for the given area. These models have high resolution to cover only a specific area. Regional models have more advantages over global models because of higher resolution to see features which are missed out on by global models. Below given are some of the highest resolution and most accurate regional models:

• Central Europe Super HD (Resolution ~1 KM)
• AROME (Resolution ~1 KM)
• ICON-D2 (Resolution ~2 KM)
• ISwiss EU 4x4 (Resolution ~4 KM)
• Euro 4 (Resolution ~4 KM)

Why do we have so many weather models?

Many national weather centres use supercomputers to run these weather models but the method of their calculation differs. Not every national model uses the same equations to solve various physical processes to determine weather patterns and they also have a difference in their resolution for collecting weather data points. Our complex atmospheric system multiple out these slight differences through time to make high variance at the end of each model output. Weather centres also control the influence of forecasting by running ensemble systems with different initial conditions and that also return different outcomes. And finally, every nation wants to quantify its standards to forecast the weather.

So, next time when you are watching a weather report on TV or browsing a weather website/application, you know that our meteorological departments are using supercomputers to run billions of calculations to forecast a complex earth's atmospheric weather system into a simplified version for our safety and planning.