ON THE VARIABLE PERFORMANCE AND CALIBRATION OF THE STEPPED FREQUENCY MICROWAVE RADIOMETER

Z. Jelenak^2,3, J. Sapp¹, S. Alsweiss¹, C. Shoup¹, S. Soisuvarn^2,3, H. Holbach^4,5,6, P. Chang³

¹GTSC, ²UCAR, ³NOAA/NESDIS/STAR, ⁴FSU, ⁵Northern Gulf Institute, ⁶NOAA/AOML/HRD

The Stepped-Frequency Microwave Radiometer (SFMR) has been a vital instrument for observing tropical cyclones, providing crucial near-real-time wind speed and rain rate data since 2004. This data has been widely utilized by operational forecasters and for satellite instrument validation.

Over the years, as more SFMR data became available from hurricane reconnaissance flights, the measurement inconsistencies began to emerge, leading National Hurricane Center forecasters to question the validity of the highest wind speed retrievals. Efforts to rectify these observed discrepancies fell short, partly due to inconclusive results from validation analyses that relied solely on dropsonde measurements.

To address these and other inconsistencies, a dedicated experiment was conducted during the 2024 hurricane season, utilizing coincident data from multiple flights with the high-resolution Imaging Wind and Rain Profiler (IWRAP) serving as a key reference. IWRAP is a double-frequency Doppler radar and scatterometer capable of providing the highest resolution atmospheric wind and reflectivity profiles in the lowest levels of the hurricane boundary layer to date, as well as ocean surface vector winds for storm structure and intensity estimates. Our analysis revealed significant inconsistencies in SFMR performance, identifying key sources of error, including: inadequate calibration, frequency drift, time-dependent measurement drift, and flaws in the forward model and inversion scheme. These issues caused variable performance both within and between flights for the same unit. To rectify discovered issues collaboration with the SFMR vendor, ProSensing, has been established in order to characterize the performance of three NOAA-owned instruments through in-lab calibration.

These findings highlight the potential for significant errors in the 20-year SFMR measurement database. In fact, a recent study on SFMR data assimilation in the HAFS project, utilizing data from the 2022-2024 hurricane seasons, found a degradation of maximum surface wind and minimum surface pressure when the data was assimilated, prompting the cessation of SFMR data assimilation during the 2025 hurricane season. This paper will discuss all findings and propose a path forward, underscoring the urgent need for improved calibration and retrieval methodologies to ensure the long-term reliability of SFMR data.