Guidance on Tropical Cyclone Peak Wind Speed Undersampling from Aircraft and Satellite Observations

Obtaining the best estimate of tropical cyclone (TC) intensity is vital for operational forecasting centers to produce accurate forecasts. Aircraft data are traditionally the standard for providing reliable information about the TC inner core and surrounding environment, but sampling strategies and observing platforms associated with these TC-penetrating aircraft have inherent deficiencies that contribute to the uncertainty of the intensity estimate. One such instrument, the stepped frequency microwave radiometer (SFMR) on the NOAA WP-3D aircraft, provides surface wind speeds along the aircraft flight track, usually sampling several times near the maximum surface wind speed. Because the design of the flight tracks tends to limit azimuthal coverage of wind speed observations, it is difficult to truly observe the maximum wind speed during a given flight. However, using a high-resolution Weather Research and Forecast (WRF) simulation of Hurricane Isabel (2003), a previous study produced SFMR-like wind speeds along simulated flight tracks. Generally, the 1-minute mean (maximum) surface winds underestimate a 6-hour running mean maximum wind (i.e. best track) by ~7.5-10%.

In this current JHT work, we make use of a high-resolution hurricane nature run (HNR1) in the place of the Isabel simulation in order to assess the underestimation of the maximum wind throughout the TC lifecycle. Additional simulations of Hurricane Bill (2009), a second hurricane nature run (HNR2), and two idealized hurricanes, which all conform to same model specifics as HNR1, are used to help understand the underestimate at various TC stages. In particular, this study examines the effect of altering the flight pattern to improve the in-storm coverage by aircraft. Underestimates are also compared to the structure of the TC, such as its size, radius of maximum wind speed, or amount of wind speed asymmetry. These factors have implications for the amount of expected underestimation, where large, asymmetric tropical storms have a much larger mean underestimate (~20%) than small, intense, and very symmetric hurricanes (~1-2%). Additionally, simulated scatterometer (ASCAT and QuikSCAT) swaths are produced to assess the underestimate by scatterometers in tropical storms and non-major hurricanes. Instrument limitations in reality prevent obtaining reliable wind speeds greater than ~40 m s^-1. Some of the impact here is due to horizontal resolution of the swath, but generally, scatterometers underestimate the peak wind speed by ~15-20%. Applying the standard rain flagging metrics also increases the underestimate. For forecast periods that include SFMR and/or scatterometer data during the 2017 Atlantic hurricane season, we apply the associated underestimate correction and compare the results to the official best track intensity.