Snapshot of AI in Enhanced Weather Forecasting Challenge
Vinko Lešić, IEEE Region 8 Vice-Chair, Member Activities
As a result of climate change, extreme weather events are becoming increasingly frequent and resulting in a growing need for more accurate real-time updated weather prediction where short term weather forecasting (nowcasting) is gaining critical importance. With availability of real-time open-source data such as Numerical Weather Prediction (NWP) forecasts, satellite and weather radar imagery, together with localized weather measurements, new and interdisciplinary possibilities are emerging in the way weather forecasts are generated.
Here, we recognize opportunity where information and communication technologies (ICT) can greatly contribute. New sensor integration technologies, based on Internet of Things (IoT), enable weather measurements available in real time but also accumulate them and form big data over the years. This enables algorithms from what we call artificial intelligence to exploit the data and create new models to support weather forecasting. Moreover, multi-modal real-time data can now be paired with machine learning approaches to improve the accuracy and reliability of regular weather predictions. We call this data fusion, and in reality, it creates hybrid models based on physics, historical data and real-time Earth observation, that we expect can improve local weather forecasts by up to 70%, depending on the specific weather variable.
The purpose of the challenge is to gather all the experts in the domains, exchange approaches and algorithms, and pinpoint guidelines towards worldwide coverage of improving the accuracy of weather forecasting. For this, we have created a competition shaped as an international hackathon with the finals being conducted in Bari, Italy, during 27-30 November 2024 as part of the IEEE International Humanitarian Technologies Conference (IHTC) 2024. Main task of the competition was to generate 7 days ahead weather forecasts, on hourly resolution, for 2 weather variables in each of the 3 case studies, which are selected as three biomes of Europe, Middle East and Africa (IEEE Region 8): Savanna Preservation, Clean Urban Air and Resilient Fields.
After a very dynamic month and nearly 10,000 views of the competition, it is now closed and teams from 10 countries have submitted their solutions. Some of those have resulted in significantly improved weather forecast compared to the conventional one that we use every day.
The top five teams are selected and those are Visionaries Team from Jordan, CUFE team from Egypt, ENEA & UNINA from Italy, Y-CAST from Germany and DERI from United Kingdom. Full list:
- Faculty of Information Technology at the Applied Science Private University in Amman, Jordan
- Cairo University, Cairo, Egypt
- Digital Environment Research Institute at Queen Mary University of London, London, UK
- University of Naples Federico and Enea Research Center, Naples, Italy
- German International University, Cairo, Egypt
The winners will present their solutions and approach on 28th November 2024 at a dedicated session during IEEE IHTC 2024. The session brings also speakers and experts from academia and industry, in the areas of artificial intelligence, meteorology and technology in climate in general. Together with other conference speakers and participants, the session will also set foundations for future stages of the competition, scaling it up and opening it to the whole world.
Savanna Preservation
The world’s famous ecosystem of grasses, shrubs, bushes, and scattered trees with open canopies, hosting our largest and most iconic animals – African savannah – is facing danger of desertification due to excessive heat and droughts as result of an increasing average temperature. Plants are perishing and no longer shelter and hold down soils, which would usually prevent the drying out and erosion of nutrients. The critical factor that rapidly speeds up this process is the increased danger of wildfires. The Savanah preservation climate challenge is focused on accurate predictions of weather variables that create favorable conditions for draughts and wildfires, to better anticipate the upcoming danger.
Clean Urban Air
Today, 55% of the world’s population lives in urban areas, a proportion that is, according to UN, expected to increase to 68% by 2050. The cities of Middle East host emphasized above-average numbers, where Bahrain as an example holds 100% of urban population. Increased global temperatures are leading to drier dry spells and hotter heatwaves, making dust on roads become airborne easier and contribute to particulate pollution, lowering the quality of the air we breathe outdoors. The US Air Quality Index, or AQI, is the system used to warn the public when air pollution is dangerous by tracking ozone and particle pollution. The Clean Urban Air climate challenge is focused on accurate predictions of weather variables that reflect the status of air quality in urban areas.
Resilient Fields
With over 17% of global production, Europe is the highest producer of wheat crops, which are the central driver and origin of our civilization. Actual and upcoming climate changes will evidently have the largest impact on agriculture crops cultivation in terms of reduced harvest, increased costs, and necessary deviation from traditional farming as our seasons are slowly becoming unrecognizable. It is expected that climate change will lower global wheat production by 1.9% by mid-century but the increasing danger of more frequent extreme weather events makes all prognoses very uncertain. The Resilient Fields challenge aims to increase the accuracy of weather forecasts for variables of significant impact to crops development, to improve the resilience of our fields by better anticipation of weather conditions.