Using SAR to Diagnose the Boundary Layer Structure of Hurricane Larry (2021)

Modern satellite C-band SAR such as Sentinel-1 can provide very high resolution measurements of the co-pol (VV) and cross-pol (VH) normalized radar backscatter cross section (NRCS) for most of the inner core of tropical cyclones. These measurements can be used to estimate km-scale surface wind vectors with reasonable accuracy up to ~80 ms^-1 and to directly measure the orientation of the rolls that routinely form in the tropical cyclone boundary layer (TCBL). The goal of this research is to use this unique data to infer the mean wind profiles within the TCBL. A NOAA NESDIS P-3 research flight was timed to under-fly a Sentinel-1 overpass of Hurricane Larry on Sep. 7, 2021. In addition to the standard instrumentation such as tail Doppler radar, flight level sensors and SFMR, the P-3 was also equipped with the IWRAP downward-looking conical scanning Ku-band radar. Hence, this flight provided an excellent dataset for advancing SAR research. The first step in the SAR analysis is to calculate the surface pressure pattern consistent with the surface wind vector field. This is done by using a reduced version of a similarity boundary layer model to estimate a corresponding pressure gradient vector for each surface wind vector. Ordinary least-squares can be used to find the best fit surface pressure pattern corresponding to the SAR image. Compared to drop sondes for a range of TCs, this method has typical RMS around 4 mb determined by comparing the measured and derived pressure differences between all possible pairs of drop sondes. For the specific Larry (2021) case, the RMS is much lower (~2 mb), although the total drop sonde count was relatively low. A field of gradient winds can be directly calculated from the pressure surface. Hence, the SAR can provide both the upper and lower boundary conditions needed by a similarity TC boundary layer model. The missing information is a reasonable estimate of the eddy viscosity profile. It is well-established that TCBL rolls are associated with large-scale shear instabilities that reach a quasi-equilibrium state. The basic characteristics of these rolls are fairly sensitive to the shape of the mean wind profiles, especially the cross-wind (largely radial) component. The roll orientation angle provides a sensitive metric that is easy to measure from SAR as the rolls impart a locally-periodic modulation of the surface wave field that manifests as a locally-periodic modulation of the NRCS. These mean wind profiles are in turn sensitive to the vertical profile of eddy viscosity. Using mesoscale model parameterizations as guide, we survey a reasonable range of eddy viscosity formulations to maximize the agreement between predicted and measured roll orientations. The final result is an estimate of the mean TCBL wind profiles. We will present comparisons of these results against high-resolution vector winds derived from IWRAP.