Evaluation of convection-allowing model ensemble members using a network of planetary boundary layer observations from the TRACER field campaign

The coastal environment is influenced by a contrast in surface temperature, roughness lengths, and planetary boundary layer (PBL) structure. Coastal breeze circulations (bay breeze (BB), sea breeze (SB), lake breeze, etc.) modulate weather conditions by advecting a maritime airmass onshore, altering air quality, and often initiating deep convection. Coastal breeze interactions with the PBL are not well simulated due to small-scale interactions being parameterized by numerical models. The choice of PBL and shortwave radiation parameterization schemes can cause large differences in SB propagation speed, depth, and onset due to variations in surface fluxes and mixing schemes. For successful numerical forecasts of air quality and local weather, coastal breeze simulations need to evolve accurately.

The TRacking Aerosol Convection interactions ExpeRiment (TRACER) field campaign collected data to understand the connections between regional breeze circulations, aerosols, and the convective cloud lifecycle in Houston, Texas. A dense dataset of PBL observations gathered across the metro has provided insight into multiscale processes. Remote sensors were placed perpendicular to the coast with increasing distance from the shore to understand the spatial variability of thermodynamic and kinematic structures. Additionally, small uncrewed aerial systems (UAS) were deployed during intensive observation periods near the coast to gather high-resolution in-situ observations in tandem with the continuous PBL profilers. Various numerical models were also used to achieve a process-oriented understanding of these complex interactions. The Warn-on-Forecast system (WoFs) is a convection-allowing model with rapidly updating data assimilation cycles designed for high-impact weather events. WoFs has 18 ensemble members with varying PBL and radiation parameterizations. The dense network of PBL observations provides a benchmark to determine the performance and potential biases accompanying various parameterizations. There is broad variability across members in the depth, arrival time, and intensity of the simulated SB case. Moreover, only a subset of members simulate the preceding BB, but none do so accurately. In turn, some ensemble members are more aggressive with CI, while, in reality, this case had very few convective cells surpassing 40 dBZ and all were short-lived. We will discuss the reasons behind these biases to offer potential solutions to improve the WoFs representation of coastal circulations and related PBL processes.