There’s a lot of talk about artificial intelligence and the benefits of its application, from dating, marketing and social media to space exploration and medical advancements. There isn’t an industry that hasn’t been affected by this dynamic tool, including the weather.
Meteorology has always struggled with the problem of big data. I would even say that science was the quintessence of big data before the word became mainstream. Due to the multivariate and chaotic nature of the weather, for more than half a century, meteorologists have processed terabytes of data and modeling variables to produce an accurate forecast. Today, we’re still processing data — now at petabyte scale — thanks to the Internet of Things, more sensors, and ensemble modeling. Writer Ted Alcorn estimates that “current (weather) models integrate approximately 100 million data elements every day, a level of complexity comparable to simulations of the human brain or the birth of the universe”.
But computing power and advances in technologies like AI have allowed us not only to analyze data faster and easier, but also to “learn” from historical data for better situational awareness and better decision making. Within the weather community, AI is being applied to several different challenges. One of the goals is to make better weather forecasts.
The forecasts are becoming more and more precise. Today, a five-day forecast has 90% accuracy, the same as a three-day forecast 25 years ago. Short-term predictions, or now projecting to hourly timeframes, are more difficult, especially due to micro-changes on the surface. Scientists from DeepMind and the University of Exeter have teamed up with the UK Met Office to build an AI-powered nowcasting system that would overcome these challenges to make more accurate short-term forecasts, including for critical storms and floods. Another research study examines the effectiveness of modeling and how AI can analyze past weather patterns to predict future events, more efficiently and accurately.
My work focus – and the area of AI that interests me particularly – is its application to predict the potential impact of weather events. The results of the weather as opposed to the weather itself.
For example, using AI in the utility industry to predict potential outages. Historical outage data is collected about a specific utility location or region and allows a computer to generate forecasts for future needs based on predicted weather conditions. He understands how infrastructure has responded to past storms, including learning differences in network hardening, realizing the age of individual infrastructure components, and maintenance practices. These datasets will provide a baseline of potential outages due to upcoming storms. The same approach can be applied with municipalities. By understanding variables such as the city’s infrastructure, topography and escape routes, as well as historical weather data, we can help cities better understand potential impact areas and safety risks to the public or infrastructure.
And, as we talk about cutting-edge technology and ideas, I think it’s important to note that the human element is always crucial to the process. A recent Wired article cited studies that found forecasts from human forecasters were more accurate than forecasts from AI.
Another area that requires human intervention is the growing need for risk communicators. They are meteorologists who push the forecast further and convey the risk or impact to a business, municipality or the public. I’ve heard several comments that when the AI is more reliable, it will be enough to toggle the weather preferences to get accurate and meaningful weather data on demand. While I agree that we will progressively have better data and predictions, I believe it will also increase the need for human experts to assess, interpret and communicate data – as well as risk and impact – in a way that has makes sense for those who need to make agile and informed decisions to protect people, infrastructure and business assets. The big question shouldn’t be human or AI forecasts, but rather how meteorologists can use improved AI to help decision-makers make the best decisions for their stakeholders.