University of Washington Study Says AI Can Improve Lightning Forecasts

Lightning is destructive force of nature and quite challenging to predict. University of Washington has done a comprehensive study and says that AI can improve lightening forecasts. Machine learning algorithms can be used to improve predictions in lightning. The University of Washington researchers say the forecasts will be helpful for preparing for potential wildfires, gives advanced safety warnings, and develop reliable long-range climate models.

The University of Washington is the first to develop machine learning algorithms and combine it with past data of lightning.  The new model brought forward by the University of Washington has combined weather forecasts with machine learning equations studying data of past events of lightning. The university presented its hybrid model at American Geophysical Union’s fall meeting.  The new technique has several advanced integrations than the existing technique.

The university through its model demonstrated that severe weather forecasts can be improved with methods based on machine learning algorithms.  The designed the systems is trained with machine learning models gathering data from 2010 to 2016. This allowed the computer based machine learning algorithms discover relationships between weather variables and lightning bolts. Further, the new technique was tested on weather from 2017 to 2019. The technique was also compared to physics based methods following actual lightning observations to evaluate the results combining both.

Several phenomena were taken into consideration such as amount of precipitation, ascent speed of storm clouds, and several other factors.   Further, the researchers hope to improve the technique using more data sources, weather variables, and other advanced techniques making its more reliable and accurate for predictions.