Development of an Optical Turbulence Neural Network Model for the Naval Surface Warfare Center Dahlgren Division (NSWCDD) Potomac River Test Range (PRTR) Utilizing Bulk Meteorological Sensors

Characterizing degradation effects of optical turbulence on High Energy Lasers (HEL) and infrared/visible imagers are integral to the system design and performance. The Naval Surface Warfare Center Dahlgren Division (NSWCDD) maintains a HEL/optical test range that includes a 2.3 km overwater open air path. Optical turbulence measurements on range are currently measured using scintillometers, differential image motion monitors (DIMM), image processing, and 3D sonic anemometers to characterize the atmospheric effects on HEL propagation. Currently, in the event of erroneous measurements or catastrophic instrumentation failure, no backup measurement exists. Additional modeling methods are needed to ensure data is continuously available to support programs testing at NSWCDD Potomac River Test Range (PRTR), which is the limitation the Naval Optical Turbulence AI/ML Model (NOTAM) for PRTR plans to address. PRTR has employed an array of bulk meteorological sensors that collect continuous surface temperature, air temperature, wind data, and moisture conditions at various locations down range, both overland and overwater. This array has been active for the past 10+ years and continues to be maintained and operated by WFHolding Incorporated. Utilizing this preexisting infrastructure, including the archive of spatially separated bulk meteorological measurements across the entire river range and an extensive path averaged optical turbulence archive, the Propagation and Measurements and Analysis Group (PMAG) at NSWCDD has developed an artificial intelligence/machine learning (AI/ML) algorithm, NOTAM, to model optical turbulence conditions for the PRTR. Due to the range’s littoral nature, the model is required to be dynamic, able to handle rapidly changing environments, and achieve enough precision to inform the test team of the optical turbulence down range. The goal of this model is to outperform the most cutting-edge physics-based surface layer models that will utilize the same inputs as the new AI/ML algorithm and validated against path averaged measurements taken on range. This presentation will examine the performance and design of the NOTAM AI/ML model.