The Joint Outdoor-indoor Large Eddy Simulation (JOULES) for Numerical Weather Prediction, Forecasting and Analysis in Urban Environments.

Large Eddy Simulation models (at scales of 1 to 5 meters) are proving to be a necessary tool for simulating and forecasting turbulence in urban areas as the aviation weather community looks to adopt the influx of Uncrewed Aerial Systems (UAS) and electric vertical take-off and landing (eVTOL) aircraft that wish to operate in these airspaces. Traditional LES models are driven by periodic lateral boundary conditions, in that, a quiescent initial condition is “spun-up” to a representative steady state of atmospheric turbulence given user specified environmental parameters and the recycling of the outflow turbulence from the downwind side of the domain to the upwind side. These types of simulations have proven useful as they allow the user to simulate a full spectrum of turbulence with a single initial atmospheric profile and can provide detailed statistical distributions of atmospheric conditions, or an ensemble realization, at discrete points in space simply by aggregating across model time. As such, the Joint Outdoor-indoor Large Eddy Simulation (JOULES) has served the community for years being run in this way. However, recent improvements to the JOULES model have enabled the implementation of an open boundary condition mode that allows for the simulation of time-varying meteorology, in particular, driving the advective tendencies of the high-resolution LES fields with a coarser resolution mesoscale weather model. Because the JOULES model is natively a GPU code and can run at faster than real-time speeds, it can be coupled to forecast mesoscale products as they become available either publicly or as they are created by a user’s forecasting pipeline. Aeris has extended this to an operational system for inline downscaling of several common NWP models (HRRR, ERA5, & WRF) to a microscale NWP forecasting service that can provide stakeholders with detailed forecast meteorology and turbulence in the urban domain at the meter scale.

Aeris will present recent results from Verification and Validation (V&V) studies that demonstrate the accuracy of JOULES compared to surface observations in environments that are densely populated with buildings, highlight significant orographic flow channeling from undulating terrain, and have flow structures that are dominated by larger scale synoptic weather patterns. Aeris will also present an analysis of the seasonal and temporal variability of urban turbulence that demonstrates the level of weather intelligence that can be obtained from an operational, cycling, microscale NWP meteorological product. Finally, Aeris will present a prototype web-based platform that allows users to setup and initiate simulations with JOULES leveraging a robust geospatial database of terrain, building, and landuse data to customize their simulation to their domain of interest.