Enhanced Marine Awareness through Real-Time Processing of Crowd-Sourced Mobile Device Observations

Accurate marine weather nowcasts and forecasts are critical for mariners to maintain situational awareness and ensure safe navigation for vessels, passengers, crew, and cargo. These projections rely on consistent observations of atmospheric and sea conditions at high spatial and temporal coverage. Unfortunately, existing approaches for marine weather observations, including fixed observational buoys, volunteer observations from trained mariners, and remotely sensed satellite data, are sparse or unreliable. Lacking consistent observations presents a challenge for generating real-time geo-specific forecasts for marine navigation.

The poster presents the development and deployment of a mobile platform (WeatherCitizen, weathercitizen.org) for collecting and distributing crowd-sourced environmental observations in a marine environment. WeatherCitizen enables collection of crowd sourced marine observations through different modes of operation, including entry of manual observations (including text, audio, and images), automated polling of internal phone sensors (e.g., barometer), and Bluetooth connectivity to external weather stations. Observational data are uploaded to a central database, filtered for quality and relevance, and pushed to nearby WeatherCitizen devices in real-time. WeatherCitizen data is aggregated with traditional marine observations and available through a web map interface or programmatic application programming interface (API).

In collaboration with Stony Brook University, we performed a field trial of WeatherCitizen on the Long Island Sound collecting observations from 10+ geospatially distributed mobile devices. Based on this dataset, we present algorithms for advanced situational awareness, including wave period, wind-speed, and assimilation of observations into forecast models. Finally, we discuss the roadmap for WeatherCitizen into the future.