Learn from your fellow PHP developers with our PHP blogs, or help share the knowledge you've gained by writing your own.
![Using AI for Weather Forecasting](https:/ /cdn.filestackcontent.com/NZd1Hj1pTvmhJqRqIYcw)
> Technology is constantly changing the way we interact, research, and react. One such way artificial intelligence is impacting our daily lives, and we may not even realize it is in weather forecasting.
The forecast we usually have been receiving in our phones and in older times primarily in newspapers, was based on data collected via satellites, radar system and weather balloons. In recent times there has been the addition of IoT based sensors as well. However, with the advent of Artificial Intelligence (AI) finding its way in numerous areas, AI has taken a role in improving the accuracy of weather as well.
### The Dataset expansion
A significantly enormous set of data is available - from the weather satellites in space, to the private and government owned weather stations which are gaining real-time data. IBM for instance has more the 0.25 million weather stations that help IBM collect real-time data. Additionally, as we are in the age of Internet of Things (IOT), each small device to big device- cellphones, solar panels and vehicles everything has become or is yet to become yet another data source. Companies like GE have installed IOT street lights, which help in monitoring air quality and humidity. These are some of the few sources which help us in collecting the vast amount of data necessary for building on the AI technology, in future these sources and the amount of available data would grow exponentially.
### Google and Weather forecast
Using the AI technology Google is able to develop a weather forecast tool, it has been trained to predict rainfalls accurately as much as six hours before. The underlying technology on which this prediction is build upon is U-Net convolutional neural network which is originally used in biomedical research. It works by taking satellite images as input and uses AI technology to transform these images into high resolution images. The only off-set is this is not real-time prediction and the delay due to complex calculations results in using six-hour old data and hence can only predict six-hours before.
### IBM and its efforts in weather prediction
The quest for IBM to venture into weather forecasting began with IBM acquiring The Weather Company. IBM plans on using the large amount of weather data available coupled with IBM Watson and the cloud platform to enhance weather forecasting. In 2019 IBM developed Global High-Resolution Atmospheric Forecasting System (GRAF) in order to forecast weather conditions 12 hours prior to a greater degree of accuracy. The radius encompassed by the GRAF is also more narrowed down up to 3 kilometers as opposed to generally being 10-15 kilometers. Another of its marvel is that it gives accurate predictions down to each hour and not just daily.
### Artificial Intelligence and Panasonic
Panasonic is the company behind TAMDAR, the weather sensor installed on commercial airplanes. With this advantage of extensive amount of data from in-flight sensors as well as publicly available data Panasonic developed Global 4D Weather. Proving to their claim of being the most advanced global forecasting platform globally they were able to timely predict Hurricane Irma in its early days.
## Uses of Weather Forecasting
Everyday life decisions are affected by weather, it makes us choose in the way we travel, things we eat and things we buy to wear. The rise in temperature may increase sales of chilled drinks, if the company is fully aware of the forecast it would be able to manage productions as per demand. AI can help brands in maximizing sales based on weather forecasts and in minimizing waste.
### Natural Disasters
The Panasonic Global 4D weather predicting Hurricane Irma is just another example where timely prediction can save millions of lives in face of situations like floods and Hurricanes. Companies like IBM combine weather forecasting data with utilities distribution network, which enables them to narrow down areas with likely outages. This enables utilities to place their workforce timely so the repair process catering to damage repairs post disasters is shortened. This in turn brings huge benefits to the overall economy.
The weather and agriculture have the most obvious correlation, each process in farming from sowing to reaping all depends on the weather. As farmers cultivate on huge farming lands, accurate information about each part of the land can help farmers in improving their crops and yield by manifolds. Weather conditions can lead to almost 90 percent of crop losses, 25 percent of these losses can be avoided using accurate AI prediction models to forecast weather and in turn improve the yield.
Sea travel has always been eventful, timely prediction of storms by using machine learning techniques and hyper-local data allows companies to plan shipments accordingly and avoid severe weather conditions that usually result in delays. Tools like IBM’s Operations Dashboard for Ground Transportation equips in enhancing productivity based on weather predictions.
Another of the implementation of AI in transportation industry corelating to weather is fuel consumption. For instance, using weather prediction models to reduce airplane fuel consumption during its ascent.
To conclude Artificial Intelligence has a key role to play in weather forecasting, weather direct or indirectly impacts each sector in the economy. As the amount of information available to improve predictions increases exponentially it gives a chance to AI to improve accuracy even further. As we continue narrowing down weather conditions precise to time and location the benefits of such advancements across all industries are innumerable.