Impacts of assimilating observations from connected vehicles into a numerical weather prediction model

Abstract Numerical weather prediction (NWP), a method of weather forecasting using equations that describe atmospheric flows and behavior, is used extensively to help predict weather conditions along roadways and has important implications for maintaining safety and efficiency on the transportation network. An insufficiently dense surface observation network presents a major limitation in obtaining the current state of the atmosphere for NWP at high spatial and temporal resolutions, such as required by road weather applications. Connected vehicle technologies, where public, private, and commercial vehicles serve as weather-observing platforms, can be used to fill in these gaps in the surface weather observation network. A pilot study was formed to quantify the impact of a dense network of observations along roadways, such as would exist with fully implemented connected vehicle technologies, on NWP. First, a simulated vehicle probe dataset was created. This dataset was assimilated into the Weather Research and Forecasting (WRF) model for select case studies, and the resulting output compared against observations and a baseline WRF run without the vehicle data assimilation. It was found that these observations had an overall positive impact on precipitation and other surface variable outputs, though in several cases the impact was slight. Major improvements occurred when using wiper status as a proxy for precipitation to apply forward error correction to a point model forecast of probability of precipitation and quantitative precipitation forecasts.