Use of Non-Traditional Observations in Weather Forecasting

We have two projects in this area , using smartphone pressures and unmanned weather-observing sailboats.

The smartphone project has been going on for roughly ten years.  Many smartphones have surprisingly good pressure sensors and such pressures are collected by several apps.  Currently, we are working with  weather.com, who app is collecting millions of pressures each day.  Our research has examined how to quality control smartphone pressure observations, and considerable progress has been made using machine learning.  We are also using these pressure in ensemble-based data assimilation systems to demonstrate that such observations can actually improve forecasts.

Smartphone Pressure Observations Over One Day Across the U.S.

A more recent project is the use of unmanned sailboats, that are constructed by a company called Saildrone.   Such saildrones are unmanned sailboats that can collect a wide range of atmospheric and ocean information, sending it back in real-time through satellites.  Currently, we are working on a project that has put a line of saildrones off the U.S. West coast, with the hope of improving U.S. weather prediction.