IMHO this is likely to travel a path similiar to the weather forecast.
I think that the alarms will get more reliable over time. Researchers will develop better models. On the consumer side there may be applications that evaluate data and forecasts to suggest appropriate actions (e. g. let me know about light shakes but wait with an alert until a treshold is reached). IoT may help to expand the sensor network and contribute to better data. I am conviced that there is a market for this.