Predictive weather data: Forecasting storm paths to protect communities
Historically, weather forecasting has been a tool used to save lives and protect property, but it typically worked along the lines of a right-or-wrong system, according to the National Center for Atmospheric Research (NCAR). In this system, predictive weather data models showed a single path a storm could potentially take. The storm then either did or did not conform to the model’s prediction, hence, the right-or-wrong single model limitation.
With predictive analytics, however, powerful computers run numerous models that consider a number of conditions, compiling and analyzing disparate data sets surrounding temperature, humidity, wind and more. These new capabilities allow forecasters to suggest a realm of paths that a storm can take and advise all communities that may be affected by the weather event. Consider several examples of how data analytics are being put to work by scientists and meteorologists that give communities everywhere a predictive edge to emergency management.
Putting predictive models to use
The key to maximizing the potential of predictive weather data is to run models more frequently and give a greater number of meteorologists access to the data. If experts can provide constructive feedback about predictions, then the accuracy and ability of future storm-path models can increase. That is, when a storm does follow a modeled path, the algorithm that created the model is supported. If it strays from the prediction, however, the algorithm can be readdressed. NCAR has taken recent steps toward this goal: weather professionals across the country can now see NCAR’s modeling results online in real time.
Scientists are looking to predict storm paths in real time by tapping into the potential of distributed data gathering. For example, if researchers can tap into apps on smartphones and other mobile devices in regions surrounding a storm event, those devices can become sources of meteorological information. This approach has recently become possible because some mobile devices are now carrying barometric sensors. When smartphones can measure shifts in conditions such as atmospheric pressure, they can feed data to evolving models as a storm progresses. Some predictive weather data labs are exploring how to collect this kind of information from users on an opt-in basis, according to an MIT Technology Review report. Ultimately, these labs could use the data to help forecast where a storm will go next.
Work is even being done with unconventional data sources such as birds, according to a recent NBC News report. Scientists recently spent a year collecting information from flocks of warblers, using microsensors attached to the animals to track their migration patterns.
The scientists found, rather unintentionally, that the birds could somehow sense tornadoes were coming days in advance, and they moved accordingly to avoid the severe weather. While the researchers are unsure how these birds sense impending storms, the knowledge could help to further improve data analytics and storm-path modeling. The behaviors that the storm-sensitive birds exhibit could eventually provide critical forecasts to predictive weather data systems, but further research is needed to unlock the secrets of how they do it.
Using storm-path predictions on the ground
Better models show where storm paths may affect communities, and similar data-driven analytics then tell emergency responders how those paths can impact buildings, people and access points. Data analytics can help local governments predict damage, prevent post-event behaviors such as crime in sectors where the electricity was cut off and allocate resources to recovery zones if the storm leaves people without shelter, food or medical support. As predictive weather data improves the accuracy of storm-path modeling, these other advanced planning functions become more effective as well.
The technology that drives predictive weather data ranges from rocket-borne weather satellites to consumers’ smartphones, but the goal is the same: create models that respond to changes in the factors that fuel storms. As these tools evolve, residents in communities where severe weather will touch down can be notified about the impending storm sooner than was previously possible. And they can proactively prepare for potential damage that may be caused by the winds, rain and flooding in the aftermath of the storm.
Attend IBM i2 Summit for a Safer Planet to learn more on how innovative and advanced human-led intelligence analysis solutions can help you develop proactive emergency plans.