Exploring the Optimal Design of a Mesoscale Boundary Layer Profiling Network for Model Forecasting Applications in Cases of Significant Severe Weather

Abstract

Rapid advancements to computational capabilities, and the resulting proliferation of high-resolution convective allowing models (CAMs), have recently revolutionized severe weather forecasting. However, CAMs are still far from perfect, and are limited by the availability of observational data. Operational observations directly above Earth's surface are relatively sparse in space and time, leaving much of the three-dimensional structure of the lower troposphere poorly resolved. This under-sampling is likely a major limiting factor to the performance of CAM forecasts, which rely on observations to inform accurate initial conditions of the atmosphere. This is especially relevant to weather phenomena that occur in the atmospheric boundary layer (ABL), where small-scale heterogeneities driven by turbulence are merely parameterized, even in the most sophisticated operational models. Numerous features and processes, some of which play an important role in the evolution of convective storms, also exist in the ABL. A mesoscale network of vertical profiling systems could provide high-resolution observations of these features in addition to the general kinematic and thermodynamic structure of the ABL. Such observations would help to fill some of the data gaps that are likely detrimental to model forecasts of severe convective storms and their hazards. This study uses observing system simulation experiments (OSSEs) to explore how assimilating simulated observations from such networks into CAMs may impact forecast performance for high-impact severe weather hazards such as tornadoes. Additionally, various components of the design and operation of a profiling network are evaluated with the intent of gaining insight into how the network should be optimally designed to provide value while minimizing cost.