Artificial intelligence, in the weather

For decades, the possibility of hitting the weather forecast depended on simple instruments: meters of physical factors such as atmospheric pressure, wind speed and direction.

Radars were added 50 years ago to detect cloud cover at a distance and soon after satellites, among other digital tools.

These devices are still in use today to capture data, but the climate prediction itself now depends on numerical models, algorithms and artificial intelligence (IA). That is why the forecast is more accurate today.

This allows you to know what will happen to the weather in the very short term. For example, the analyzes without AI gave the experts the possibility to know if it would rain in the following days; but not what would happen in the following hours or minutes, nor the magnitude of the events.

According to Scientific American magazine, AI has a lot of potential in meteorology. Techniques such as ‘machine learning’ can be used in applications for the analysis of serial data, trends in meteorological variables and everything that involves the use of large volumes of data.

Although the benefits are obvious, it also shows that AI has limitations when it comes to predicting the weather. This because it depends on factors such as: quality of the information collected; the available computing capacity; the level of knowledge of programming and numerical modeling. It also reveals that the effectiveness of AI is subject to the place in the world where the forecast is intended. Thus, Ecuador has more variables to consider than other parts, because it is crossed by a mountain range, it is located on the shores of the sea, it has a rainforest and it is in a tropical zone in the middle of the planet, exposed to a large amount of solar radiation.

READ  Black June: Brexit, wave of refugees, euro disaster How the re-establishment of Europe succeeds

Vladimir Arreaga, specialist in weather forecasting National Institute of Meteorology and Hydrology (Inamhi), says that the use of artificial intelligence is crucial due to the number of variables that must be analyzed. A good

An example is Quito: in very short periods the sun shines, there are torrential rains, hail falls and the sky clears again.

Another characteristic of Ecuador is the presence of different microclimates and close to each other. This is seen when it rains more in the southern part of the capital than in the north of the city. The AI achieves precision in these scenarios by making the numerical models fit local reality.

Mario Morales, an expert in artificial intelligence systems, states that climatology is the ideal branch to apply his specialty. AI allows you to automatically analyze data regardless of sample size. They can even use algorithms able to learn to predict the weather based on historical data for an area.

Arreaga details that, in fact, the Inamhi already uses numerical models for weather forecasting. The institution has its own area of ​​mathematical modeling.

The entity is getting ready to apply an AI integration project to these numerical meteorological models. It will be done in the section Inamhi Climate Prediction.

With this integration, greater reliability will be achieved when running the numerical models locally. Arreaga says that these have traditionally been run to global level. This meant that they had a resolution that was too broad and not very specific for the country.

However, the pandemic and the economic crisis have caused delays in the project. Arreaga recalls that to achieve an accurate prediction with the new technology, inputs are needed, such as a greater number of meteorological stations for the data collection. At this time, on the contrary, some of these stations have been closed due to the crisis.

READ  Anthony Joshua, Beth England, Tao Geoghegan Hart and more heroes of the capital - EzAnime.net

Currently, the data obtained in the country are integrated into numerical analyzes every three hours. The information is sent abroad, to the climate platforms of Washington, in the United States, and São Paulo, in Brazil. This is how they are incorporated into global data. Satellite information is added to these parameters. Then, the processed data returns to the country for a new local numerical run.

The AI ​​will avoid the entire journey and the forecast will return faster to Ecuador.

.

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.