Reconciling the NHC Intensity Estimates and Observed Wind Speeds for 2017 Hurricanes Harvey and Irma

The 2017 hurricane season has afforded us some important opportunities to observe land-falling hurricanes in well instrumented regions of the United States. In this paper, we describe the landfall surface observation datasets for Hurricanes Harvey and Irma, and reconcile the intensity estimates in the National Hurricane Center’s (NHC) published tropical storm advisories with the maximum wind speeds observed around landfall. When the over-water storm estimate and inland observing environments are reconciled, we find that the differences are not as great as they initially appear. This has important ramifications for both public safety, the insurance industry and others since there is a widely-held belief that the NHC often over estimates storm intensity by a category or more.

After a period of rapid intensification, Hurricane Harvey made landfall near Rockport in south-western Texas at its peak intensity of a Category 4 hurricane. The main story of Harvey is its very slow movement in the four days post its Rockport landfall, and the record breaking rainfall that occurred around the Houston area that resulted. However, the destructive winds during landfall are also a remarkable story. As the storm approached shore, the eyewall structure around Harvey was asymmetrical and, consistent with the radar imagery at landfall, the strongest surface winds were observed in the leading edge and front-left quadrant of the eyewall. There may have also been equally significant winds in the front-right quadrant as is commonly expected, however, there were few observations in this region right at landfall.

After passing over the central Keys, Hurricane Irma made landfall over Marco Island in south-western Florida as a Category 3 hurricane. Irma was a Category 4 storm in the hours leading up to landfall, although its eyewall structure remained less organized than had been exhibited prior to interacting with Cuba. This left Irma with significant variations in the wind field within the land-falling eyewall with the strongest winds appearing to be in the leading edge and front right quadrant of the eyewall.

Fortunately, the land-falling cores of both storms missed major metropolitan areas. As such, they also missed the densest portions of the state and federal observing systems. Compounding the relative lack of publically available observing system data were the system outages that were common as the high winds approached. These outages resulted from a mix of communication and power failures as well as from physical damage to the observing systems themselves. The generally more robust data from the local and/or private mesonets of WeatherFlow, Earth Networks and TCOON filled in many of the data gaps. These were further complimented with the insertion of the deployable assets from the University of Florida (FCMP Towers), Texas Tech University (StickNet), the Center for Severe Weather Research and the University of Oklahoma into the biggest of the remaining gaps, where impacts were expected to be the greatest. Together these observations provide us with comprehensive datasets and great insight into the wind fields at landfall.

At a glance, the maximum wind speed observations appear to be a category or more (10‑20 mph) below the landfall intensity estimated by the NHC. However, if we look closely at how the NHC intensities are derived, and correctly interpret the highly-localized surface observations, standardizing both to the same reference conditions, we find that there is a very good match. Specifically, for the observations, we consider the anemometer height, upwind fetch, instrument response characteristics, and data acquisition methods, and for the NHC intensity estimates we consider the contributing observations and operational assessment protocols. Understanding the relationship between the wind speeds stated in tropical cyclone advisories and those observed is critical to improving emergency preparedness and response, the communication of risk to the media and general public, and to risk assessments made by the insurance and reinsurance industries.